Virtual power plant user classification method and system based on iterative data analysis

Abstract

The invention discloses a virtual power plant user classification method and system based on iterative data analysis. The method comprises the following steps: extracting a power consumption behavior characteristic index of a user based on historical power consumption information of a virtual power plant user, the power utilization behavior characteristic indexes comprise an average load index, an average daily peak-valley difference rate index, a daily load rate index, a load fluctuation rate index, a weekly load rate index, a peak-valley average electricity price index, a peak shifting level index and a carbon emission level index. And constructing a comprehensive evaluation matrix based on the standardized indexes, and calculating the weight of each power consumption behavior characteristic index by adopting an entropy weight method according to the comprehensive evaluation matrix. And based on the power consumption behavior characteristic indexes and the corresponding weights, classifying virtual power plant users by using an ISODATA clustering algorithm. According to the scheme, quick response and accurate classification of user types are realized.

Description

Technical Field

[0001] The present invention belongs to the field of virtual power plants, and in particular relates to a virtual power plant user classification method and system based on iterative data analysis. Background Art

[0002] Virtual power plants are power coordination management systems that achieve aggregation and coordinated optimization of massive distributed new energy, energy storage systems, controllable loads, electric vehicles, etc., participate in power grid operation and power market, and will be more competitive in the future power market. The power grid used by virtual power plants to provide services to power users needs to provide high-quality and stable electricity to meet the needs of each user for electricity in production and life. However, different power users have different ways of using electricity, and the problems caused by the lack of coordination between power supply and demand have also begun to intensify. For example, the gradual increase in the peak-to-valley difference in electricity usage habits, the continuous increase in the maximum load, and the reduction in the number of maximum load utilization hours in a year will increase the difficulty of peak-to-peak regulation of virtual power plants. Therefore, it is necessary to establish a scientific and reasonable operation and regulation strategy. The uncertainty of load fluctuations on the demand side will also affect the safety and certainty of user response. In the existing technology, virtual power plants cannot comprehensively consider factors such as consumers' actual power consumption, willingness to participate, response ability, incentive costs, and changes in the value of the power market, that is, they cannot accurately classify users. In addition, the existing clustering algorithms find it difficult to divide the load curve of the virtual power plant when the user scale is large, the data volume is large, and the data dimension is high; the selection of the initial cluster center will also have a great impact on the convergence rate of the load curve clustering algorithm, and the use of Euclidean distance to calculate the distance of the load curve lacks rationality. Summary of the invention

[0003] In order to solve the deficiencies in the prior art, the present invention provides a method and system for classifying users of a virtual power plant based on iterative data analysis to solve the technical problem of user classification of a virtual power plant.

[0004] In order to solve the above technical problems, the present invention adopts the following technical solutions.

[0005] The present invention first discloses a virtual power plant user classification method based on iterative data analysis, the method comprising the following steps: Step 1: Extract the user's electricity consumption behavior characteristic index based on the historical electricity consumption information of the virtual power plant user, wherein the electricity consumption behavior characteristic index includes the average load index, the average daily peak-to-valley difference rate index, the daily load rate index, the load fluctuation rate index, the weekly load rate index, the peak-to-valley average electricity price index, the peak-shifting level index and the carbon emission level index.

[0006] Step 2: construct a comprehensive evaluation matrix based on the standardized index, and calculate the weight of each power consumption behavior characteristic index using the entropy weight method according to the comprehensive evaluation matrix.

[0007] Step 3: Based on the electricity consumption behavior characteristic index and the corresponding weight, the virtual power plant users are classified using the ISODATA clustering algorithm.

[0008] The present invention further includes the following preferred embodiments: The average load index That is, the average user load: , in, is the load value of user n at the dth sampling time; D is the sampling point in the detection cycle.

[0009] The peak-valley average electricity price index It represents the electricity price usage time of user n during the detection process, and is calculated based on the weighted average of the load value: , in, is the value of the time-of-use electricity price at the dth sampling moment; Indicates the conversion factor of power-electricity, the unit is h.

[0010] The average daily peak-to-valley rate index is the average daily value of the difference between the peak and valley values ​​of user load curve n: , in, , Indicates that user n is in The maximum load value during the day, Indicates that user n is in The minimum load value of the day, in kW; is the total number of days in the sampling cycle; Indicates the number of sampling points in a day, .

[0011] The daily load factor index The load curve of user n in the sample period represents the average value of the overall load rate in one day: .

[0012] The load fluctuation index It represents the ratio between the load standard deviation and the load average of user n during the sampling period: .

[0013] Weekly Load Factor Index Indicates the ratio of the average daily power consumption of user n in the most recent week within the sampling period to the maximum power consumption: , in, represents the electricity consumption of user n on day φ; is the maximum daily electricity consumption of user n in that week.

[0014] The peak level index It is expressed as the average absolute value of the difference between the user load curve normalized by user n during a sampling period and the local total load curve: , Where y(x) is the normalized function of the load curve: , Indicates the total load size of the area where the user is located at the dth sampling time; The carbon emission level index It represents the average value of the absolute value of the difference between the user load curve and the baseline load curve of user n during the sampling period: , in is the normalized function of the load curve: , Indicated in The user baseline load at a sampling moment.

[0015] The present invention also discloses a virtual power plant user classification system based on iterative data analysis using the aforementioned virtual power plant user classification method based on iterative data analysis, comprising: The electricity consumption behavior characteristic index extraction module is used to extract the user's electricity consumption behavior characteristic index based on the historical electricity consumption information of the virtual power plant user. The electricity consumption behavior characteristic index includes the average load index, the average daily peak-to-valley difference rate index, the daily load rate index, the load fluctuation rate index, the weekly load rate index, the peak-to-valley average electricity price index, the peak-shaving level index and the carbon emission level index.

[0016] The power consumption behavior characteristic index weight calculation module is used to construct a comprehensive evaluation matrix based on the standardized index, and calculate the weight of each power consumption behavior characteristic index using the entropy weight method according to the comprehensive evaluation matrix.

[0017] The user classification module is used to classify virtual power plant users based on the electricity consumption behavior characteristic index and the corresponding weight using the ISODATA clustering algorithm.

[0018] Accordingly, the present application also discloses a terminal, including a processor and a storage medium; The storage medium is used to store instructions; The processor is used to operate according to the instructions to execute the steps of the aforementioned virtual power plant user classification method based on iterative data analysis.

[0019] Correspondingly, the present application also discloses a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the aforementioned virtual power plant user classification method based on iterative data analysis.

[0020] The beneficial effect of the present invention is that, compared with the prior art, the present invention provides a virtual power plant user classification method and system based on iterative data analysis, which achieves rapid response and accurate classification of user types by integrating the multi-dimensional electricity consumption behavior characteristics of emerging market entities into the classification method. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is the multi-dimensional user portrait diagram in the present invention.

[0022] Figure 2 It is the execution flow chart of the ISODATA clustering algorithm in the present invention.

[0023] Figure 3 It is a load data example diagram in the present invention.

[0024] Figure 4 It is a comprehensive load curve diagram of the user in the present invention.

[0025] Figure 5 It is a normalized user characteristic index distribution diagram in the present invention.

[0026] Figure 6 It is the total load baseline diagram of 150 users in the present invention.

[0027] Figure 7 It is the clustering result in the present invention. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solution and advantages of the present invention more clear, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0029] The embodiments described in this application are only some embodiments of the present invention, not all embodiments. Based on the spirit of the present invention, other embodiments obtained by ordinary technicians in this field without creative work are all within the protection scope of the present invention.

[0030] In view of the shortcomings of the prior art, the present invention proposes a virtual power plant user classification method and system based on iterative data analysis, which analyzes the electricity consumption characteristics of the current emerging market players from multiple dimensions, and identifies the characteristics of different emerging market players. By constructing an emerging market player electricity consumption behavior index system and calculating the weights, the characteristics that affect the electricity consumption behavior of emerging market players are identified.

[0031] The virtual power plant user classification method based on iterative data analysis disclosed in the present invention comprises the following steps: Step 1: Extract the user's electricity consumption behavior characteristic index based on the historical electricity consumption information of the virtual power plant user, wherein the electricity consumption behavior characteristic index includes the average load index, the average daily peak-to-valley difference rate index, the daily load rate index, the load fluctuation rate index, the weekly load rate index, the peak-to-valley average electricity price index, the peak-shifting level index and the carbon emission level index.

[0032] In order to maximize the mobilization of funds and efficiency, the actual electricity consumption of users, participation willingness, reflection ability, incentive cost, and market price changes should be fully considered and analyzed. Based on the measurement of the power consumption information system, the characteristics of power consumption behavior are studied. Based on the traditional impact index, the present invention defines average load, peak-to-valley difference, daily load factor, load fluctuation rate, weekly load rate, and peak-to-valley average electricity price, peak shifting level, carbon emission level, etc. to describe the load curve, as shown in Table 1.

[0033] Table 1

[0034] The calculation method of each index is as follows: The average load index That is, the average user load, in kW: , in, is the load value of user n at the dth sampling time, in kW; D is the sampling point in the detection cycle. This index reflects the total energy consumption of the user and sets the upper and lower limits of the index.

[0035] The peak-valley average electricity price index It represents the electricity price usage time of user n during the detection process, which is calculated based on the weighted average of the load value. The unit is yuan / kWh. The calculation formula is as follows: , in, is the value of the time-of-use electricity price in yuan / kWh at the dth sampling moment; Indicates the conversion factor of power-electricity, in h (for example, if the sampling frequency is 24 o'clock every day, =1h).

[0036] The average daily peak-to-valley rate index is the average daily value of the difference between the peak and valley values ​​of user load curve n: , in, , Indicates that user n is in The maximum load value during the day, Indicates that user n is in The minimum load value of the day, in kW; is the total number of days in the sampling cycle; Represents the number of sampling points in a day, thus obtaining .

[0037] The daily load factor index The load curve of user n in the sample period represents the average value of the overall load rate in one day: , The load fluctuation index It represents the ratio between the load standard deviation (kW) and the load average (kW) of user n during the sampling period: , Weekly Load Factor Index Indicates the ratio of the average daily electricity consumption (kWh) to the maximum electricity consumption (kWh) of user n in the most recent week during the sampling period: , in, represents the electricity consumption of user n on day φ, in kWh; The maximum daily electricity consumption of user n in the week, in kWh.

[0038] The present invention uses the following four indexes to describe the load curve of the user's average energy consumption: average daily peak-to-valley difference index , daily load factor index , Load Fluctuation Index , Weekly Load Factor Index If the user's energy consumption is stable, it is not common, indicating that the user has a strong ability to control and manage his own load. On the contrary, the user's load curve will have a peak in electricity consumption at an instant, which is similar to the peak of household energy consumption, that is, the midday peak and the late peak overlap slightly. And it should be noted that the peak of the total current curve should be formed as part of the load aggregator.

[0039] The peak level index It is expressed as the average absolute value of the difference between the user load curve normalized by user n during a sampling period and the local total load curve.

[0040] , Where y(x) is the normalized function of the load curve: , It indicates the total load size in the area where the user is located at the dth sampling moment, in kW. This index can reflect the similarity between the user-side and regional-side power grids. Since the regional load curve represents the total electricity consumption in a region, when real-time prices are implemented, the real-time prices can reflect the supply and demand relationship in this region. When the user's maximum change level table is close to 0, it means that energy consumption may be higher during peak hours, so it is considered a priority for integration into demand response projects.

[0041] The carbon emission level index It represents the average value of the absolute value of the difference between the user load curve of user n and the baseline load curve during the sampling period.

[0042] , in is the normalized function of the load curve: , Indicated in The user's benchmark load at a sampling moment, in kW. This index reflects the deviation between the actual output and the benchmark during the user's request response.

[0043] Step 2: construct a comprehensive evaluation matrix based on the standardized index, and calculate the weight of each power consumption behavior characteristic index using the entropy weight method according to the comprehensive evaluation matrix.

[0044] First, the original electricity consumption data is normalized. The normalization function As shown below: , in, Represents the original electricity consumption data, Expressed as the maximum value of the initial sequence; It is expressed by the minimum value of the initial sequence; is the maximum spacing after standardization; is the minimum spacing after normalization.

[0045] On this basis, the entropy weight method is used to assign weights to the power behavior characteristics of power users. of The standardized index of each item is used to generate a comprehensive evaluation matrix of the sample users. : , in, are the m standardized indexes of user n. Based on the comprehensive evaluation matrix E, the entropy of each user's power demand behavior characteristic index is calculated : , In special cases, when season .

[0046] Therefore, the above formula calculates the weight value of each electricity consumption behavior characteristic index: : , Step 3: Based on the electricity consumption behavior characteristic index and the corresponding weight, the virtual power plant users are classified using the ISODATA clustering algorithm.

[0047] The present invention uses a distance measurement method to evaluate the behavior of users in the power system, that is, the user's electricity usage habits are used as dimensional vector to calculate two different The distance between dimensional vectors. The greater the distance, the more different the electricity usage habits of the two users are. The closer the distance, the more similar the electricity usage habits of the two users are. Finally, similar users are classified into the same group.

[0048] First, pre-set the control parameters. Initialize the number of predicted load curve cluster centers K. The minimum number of users in any cluster center is , the standard deviation threshold of the load curve distance distribution of basic users in the same cluster center is , the minimum distance threshold between two cluster centers is , combining the minimum distance thresholds of the two cluster centers, after one iteration, the maximum divisor of the clusters obtained is , Iterations .

[0049] The specific steps of the algorithm are: Step (1): Response to the request User-built benchmark load matrix : , in ; Collect historical user load data and record them in the form of a matrix, where each element of the matrix represents the power load of a specific user in a specific time period; The predicted benchmark load is obtained by calculating the average value of the historical load data, and the predicted benchmark load is adjusted. The adjustment coefficient can be obtained by comparing the load data at a preset time before the event moment with the load at the same time period of the predicted benchmark load; Step (2): Random selection Initial cluster centers ( may be different from K); Step (3) Calculate the Euclidean distance from each user to the center of each cluster based on the sample user's benchmark load information (e.g., user load history data, load curve, load forecast data, user power consumption behavior pattern, power consumption basic load, etc.), and then assign them to the cluster closest to them. , until the user distribution is completed; Step (4) Judgment Sample size and predefined thresholds If , the cluster To remove, the number of cluster centers is reduced by 1; Step (5) centrally updates a single cluster and calculates the average load of the users who require response: , in, for The number of baseline load curves in .

[0050] Step (6) Calculate each cluster The load curve of the sample is sampled to the center of each cluster The average distance And the total average distance of all sample curves to the center of their cluster : , , (7) Status determination: If the current number of iterations reaches the set maximum value, as well as Set to , if the number of iterations is 0, jump to step (6); When the current iteration number is an even number and the number of clusters is less than 2 times the preset value, the method proceeds to step (8); otherwise, jumps to step (6); When the number of clusters is less than half of the preset value, the method proceeds to step (8).

[0051] (8) Measure the maximum value of the cluster center standard deviation of the corresponding branch distance of the sample load curve in each cluster , when the maximum value exceeds the threshold When the following relationship is satisfied, the cluster center is split into two new cluster centers: , Generate new cluster centers based on the corresponding components of the original cluster centers: , If the split is successful, return to step (3).

[0052] Calculate the distance between all cluster centers : , N c is the number of cluster centers, s is the number of iterations; Will Minimum distance threshold The value of of Arrange in ascending order, starting with the smallest first item. First, merge the two cluster centers and , calculate the new center by the following formula : , N x The newly generated h x The number of cluster centers, S x h x The distance between cluster centers; N x h d The number of cluster centers, S d h d The spacing between cluster centers.

[0053] Starting with the next item in the increasing sequence, check the relevant categories and perform the above formula for the categories that are not already included by both elements, as well as each aggregation correction , until the distance increasing sequence is traversed.

[0054] If the current iteration is not followed and the input hypothesis parameters have not changed, then return to step (2); if the current iteration is the last iteration, then the process ends. If the input parameters change, then re-cluster. The clustering process is summarized as follows Figure 2 shown.

[0055] In order to make the solution of the present invention clearer, it is further described below in conjunction with specific experimental examples.

[0056] The load information used in the experimental analysis example is the 24-hour load curve of multiple users in the same area of ​​a province between August 14, 2023 and August 21, 2023, and the total load curve of the same period and the same granularity. The total regional load curve as of August 21 is as follows: Figure 3 As shown. It can be seen that the weekday summer load curve in this area has a "double peak" feature, with smaller energy consumption peaks at 11 a.m. and 8 p.m. to 10 p.m., and relatively lower load levels in the early morning. The peak-valley load user and the residential user signed a price package for the peak period from 9:00 a.m. to 11:00 p.m., with a maximum price of 1.095 yuan / kWh. From 9:00 a.m. to 11:00 p.m., each kWh is 0.515 yuan. Using historical data, understand the user's investment in the demand reflection project and establish an average load index . In the screening of 0.8kW and above, there are 150 certified users. As of August 21, Figure 4 The complete load curve for these 150 users is shown.

[0057] The total load curve of 150 users is obviously similar to the total load curve of the area, and also has the characteristics of "double peaks". The peak of the utilization curve at noon appears from 11 am to 12 noon, and the peak at night appears at 9 pm, indicating that the operation of users in this area has improved the overall load level to a certain extent, meeting the needs of users for peak regulation and valley filling of the total load curve.

[0058] Based on the load data of all users in the test sample between August 14, 2023 and August 21, 2023, the electricity consumption behavior index of each user can be calculated: Peak-valley average electricity price index , Average daily peak-to-valley difference index , daily load factor index , Load Fluctuation Index , Weekly Load Factor Index and peak level index , and after normalization according to the normalization function, the calculation result is as follows Figure 5 As shown, the horizontal axis represents the number of users and the vertical axis represents the value of the normalized index.

[0059] After determining the characteristic index of the power consumption characteristics, they are weighted to obtain the specific values ​​of each index, as shown in Table 2.

[0060] Table 2

[0061] like Figure 4 As shown in Table 2, the average daily peak-to-valley difference index Compared with the load volatility index, its dispersion is the smallest and has the smallest weight, which is 0.0579; the results show that the load volatility index The data has a low degree of dispersion, but a high degree of significance, which is 0.3186. In some exceptional cases, if a metric value is the same for all sampled users, then the metric value is 0.

[0062] like Figure 5 As shown, the average method calculates the average load curve of the days before the request response date, determines the date when the user achieves load control, filters the baseline load, and takes the total load of 150 users as the baseline.

[0063] Figure 6 It shows that the total reference load of the other 150 users is Figure 4 The total load curves measured by the users in the table are very similar and also show the characteristics of "double peaks", that is, the peak load occurs from 11 am to 12 pm and from 9 pm to 10 pm. Table 3 shows the cluster analysis of the baseline load of the 150 users and explains the parameter settings of ISODATA clustering.

[0064] Table 3

[0065] The clustering results are shown in the attached Figure 7 According to the clustering results, there are 68 frequency-based emerging market entities, 51 energy-based emerging market entities, and 31 power-based emerging market entities. Table 4 shows the centers of the clusters.

[0066] Table 4

[0067] It can be seen that the ISODATA clustering method can reasonably classify which type of emerging market entity the user belongs to.

[0068] The beneficial effect of the present invention is that, compared with the prior art, the present invention provides a virtual power plant user classification method and system based on iterative data analysis, which integrates the multi-dimensional electricity consumption behavior characteristics of emerging market entities into the classification method, thereby achieving rapid response and accurate classification of user types. Through experimental analysis, emerging market entities are aggregated into frequency-type, power-type and energy-type users to participate in demand response, verifying the effectiveness and feasibility of the method of the present invention.

[0069] The present invention may be a system, a method and / or a computer program product. The present invention also discloses a virtual power plant user classification system based on iterative data analysis based on the aforementioned virtual power plant user classification method based on iterative data analysis, comprising: The electricity consumption behavior characteristic index extraction module is used to extract the user's electricity consumption behavior characteristic index based on the historical electricity consumption information of the virtual power plant user. The electricity consumption behavior characteristic index includes the average load index, the average daily peak-to-valley difference rate index, the daily load rate index, the load fluctuation rate index, the weekly load rate index, the peak-to-valley average electricity price index, the peak-shaving level index and the carbon emission level index.

[0070] The power consumption behavior characteristic index weight calculation module is used to construct a comprehensive evaluation matrix based on the standardized index, and calculate the weight of each power consumption behavior characteristic index using the entropy weight method according to the comprehensive evaluation matrix.

[0071] The user classification module is used to classify virtual power plant users based on the electricity consumption behavior characteristic index and the corresponding weight using the ISODATA clustering algorithm.

[0072] Based on the spirit of the present invention, those skilled in the art can easily think that a computer program product can be obtained based on the aforementioned virtual power plant user classification method based on iterative data analysis. The computer program product may include a computer-readable storage medium, which carries computer-readable program instructions for enabling a processor to implement various aspects of the present disclosure. That is, the present application also includes a terminal, including a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of the aforementioned virtual power plant user classification method based on iterative data analysis.

[0073] Computer readable storage medium can be a tangible device that can keep and store the instructions used by the instruction execution device. Computer readable storage medium can be, for example, - but not limited to - electrical storage device, magnetic storage device, optical storage device, electromagnetic storage device, semiconductor storage device or any suitable combination of the above. More specific examples (non-exhaustive list) of computer readable storage medium include: portable computer disk, hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disk read-only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, mechanical encoding device, such as punch card or groove protrusion structure with instructions stored thereon and any suitable combination of the above. Computer readable storage medium used here is not interpreted as instantaneous signal itself, such as radio wave or other free propagating electromagnetic wave, electromagnetic wave propagated by waveguide or other transmission medium (for example, light pulse by optical fiber cable) or electrical signal transmitted by wire.

[0074] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0075] The computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as "C" language or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect through the Internet). In some embodiments, by using the state information of the computer-readable program instructions to personalize an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA) or a programmable logic array (PLA), the electronic circuit may execute the computer-readable program instructions, thereby implementing various aspects of the present disclosure.

[0076] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A virtual power plant user classification method based on iterative data analysis, characterized in that: The following steps are involved: Step 1: Extract the user's electricity consumption behavior characteristic index based on the historical electricity consumption information of the virtual power plant user, and the electricity consumption behavior characteristic index includes the average load index, the average daily peak-to-valley difference rate index, the daily load rate index, the load fluctuation rate index, the weekly load rate index, the peak-to-valley average electricity price index, the peak-shifting level index and the carbon emission level index; the average load index is the average user load : ， in, is the load value of user n at the dth sampling time; D is the sampling point in the detection cycle; Step 2: construct a comprehensive evaluation matrix based on the standardized index, and calculate the weight of each power consumption behavior characteristic index using the entropy weight method according to the comprehensive evaluation matrix; Step 3: Based on the electricity consumption behavior characteristic index and the corresponding weight, the virtual power plant users are classified using the ISODATA clustering algorithm.

2. The virtual power plant user classification method based on iterative data analysis according to claim 1 is characterized in that: The peak-valley average electricity price index It represents the electricity price usage time of user n during the detection process, and is calculated based on the weighted average of the load value: ， in, is the value of the time-of-use electricity price at the dth sampling moment; Indicates the conversion factor from power to electricity.

3. The virtual power plant user classification method based on iterative data analysis according to claim 2 is characterized in that: The average daily peak-to-valley rate index is the average daily value of the difference between the peak and valley values ​​of user load curve n: ， in, , Indicates that user n is in The maximum load value during the day, Indicates that user n is in The minimum load value of the day, in kW; is the total number of days in the sampling cycle; Indicates the number of sampling points in a day, .

4. The virtual power plant user classification method based on iterative data analysis according to claim 3 is characterized in that: The daily load factor index The load curve of user n in the sample period represents the average value of the overall load rate in one day: 。 5. The virtual power plant user classification method based on iterative data analysis according to claim 4 is characterized in that: The load fluctuation index It represents the ratio between the load standard deviation and the load average of user n during the sampling period: 。 6. The virtual power plant user classification method based on iterative data analysis according to claim 5 is characterized in that: Weekly Load Factor Index Indicates the ratio of the average daily power consumption of user n in the most recent week within the sampling period to the maximum power consumption: , in, Indicates that user n is in Daily electricity consumption; is the maximum daily electricity consumption of user n in that week.

7. The virtual power plant user classification method based on iterative data analysis according to claim 6 is characterized in that: The peak level index It is expressed as the average absolute value of the difference between the user load curve normalized by user n during a sampling period and the local total load curve: ， in is the normalized function of the load curve: , Indicates the total load size of the area where the user is located at the dth sampling time.

8. The virtual power plant user classification method based on iterative data analysis according to claim 7 is characterized in that: The carbon emission level index It represents the average value of the absolute value of the difference between the user load curve and the baseline load curve of user n during the sampling period: , Normalization function of the load curve: , Indicated in The user baseline load at a sampling moment.

9. A virtual power plant user classification system based on iterative data analysis, characterized in that: include: The power consumption behavior characteristic index extraction module is used to extract the power consumption behavior characteristic index of the user based on the historical power consumption information of the virtual power plant user. The power consumption behavior characteristic index includes the average load index, the average daily peak-to-valley difference rate index, the daily load rate index, the load fluctuation rate index, the weekly load rate index, the peak-to-valley average electricity price index, the peak-shifting level index and the carbon emission level index; the average load index is the average user load : , in, is the load value of user n at the dth sampling time; D is the sampling point in the detection cycle; A power consumption behavior characteristic index weight calculation module is used to construct a comprehensive evaluation matrix based on the standardized index, and calculate the weight of each power consumption behavior characteristic index using an entropy weight method according to the comprehensive evaluation matrix; The user classification module is used to classify virtual power plant users based on the electricity consumption behavior characteristic index and the corresponding weight using the ISODATA clustering algorithm.

10. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is used to operate according to the instructions to execute the steps of the virtual power plant user classification method based on iterative data analysis according to any one of claims 1-8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the virtual power plant user classification method based on iterative data analysis described in any one of claims 1 to 8 are implemented.

Images