Regional demand side multi-element load clustering method and system

Abstract

The invention discloses a regional demand side multi-element load clustering method and system. The method comprises the steps of collecting multi-element load historical data in typical seasons of a park and temperature data and holiday and festival information on corresponding time nodes; calculating a historical mean value of the multi-element loads, judging the authenticity of the data, and preprocessing the historical data of the multi-element loads to obtain preprocessed historical data of the multi-element loads; adopting a CH-K-means clustering method to perform multi-element load curve clustering on the preprocessed multi-element load historical data to obtain a clustering result; and multi-element load distribution is formed. The multi-element load clustering method is an innovative regional demand side load management method, and aims to perform accurate classification and optimal aggregation according to different types of load characteristics so as to realize more effective energy utilization and load scheduling. By performing clustering analysis on the load data, the method can identify and distinguish the load types of different energy consumption modes, and aggregate to form multi-element load distribution, thereby providing detailed basic data for intelligent load management.

Description

Technical Field

[0001] The present invention relates to the field of energy, and in particular to a method and system for clustering multi-loads on a regional demand side. Background Art

[0002] In the energy field, load refers to the instantaneous power of users using different types of energy such as cold, heat, and electricity, which has time-series correlation and dynamic characteristics. Different types of users usually have personalized load characteristics. For this reason, the starting point and foothold of load research is load profiling, that is, exploring the diverse and personalized characteristics of loads from different angles. Grasping the quantitative relationship between multiple loads is not only conducive to subsequent demand-side load forecasting and interactive control, but also of great benefit to the design and operation of multi-energy cogeneration systems on the supply side.

[0003] There are many factors that affect the cold, hot and electric loads on the demand side of the campus energy Internet, which can be roughly summarized into three aspects: inherent attributes, operational attributes, and socio-economic attributes. Intrinsic attributes mainly refer to the relatively stable physical and spatial characteristics of end users. For building users, they generally include building orientation, enclosure structure, window-to-wall ratio, etc. The inherent attributes have a great influence on the heating, heat transfer and heat dissipation characteristics of the building, thereby affecting the scale of cold and hot loads. Operational attributes mainly refer to process production processes, personnel work and rest time, climate and meteorological conditions, etc. The former will affect the timing characteristics of the load, and the latter is the decisive factor of cold and hot loads. Socio-economic attributes are the environment in which the regional integrated energy system is located, such as time-of-use electricity prices, tiered electricity prices, government regulation policies, etc.

[0004] The typical RNN model has a memory function, but as the time depth increases, the memory error will be amplified, resulting in gradient disappearance or gradient explosion, making it impossible for the RNN to continue to learn information with too large a time series. The load base on the demand side is large and complex, and the clustering results are affected by many factors in the extraction process. It is necessary to conduct preliminary experimental work when selecting the clustering method to ensure good matching effects in all links. At the same time, in view of the multivariate and multi-factor coupling characteristics, it is also necessary to optimize and improve the traditional clustering algorithm to meet the needs of rapid extraction of large-scale data.

[0005] Regarding the regional demand side multi-load clustering method, the existing scheme has the following disadvantages:

[0006] The application number is CN201911199324.1, and the name is a method and system for clustering power loads. The present invention discloses a method and system for clustering power loads, and relates to the field of power system automation, and specifically to a method and system for clustering power loads. By obtaining the load time series of each user and inserting a phase shift into the load time series of each user, an embedded sequence is obtained, and a linear neural regression method is used to generate mapping function parameters for the embedded sequence, and the users are clustered using the similarity of the mapping function parameters of each user to obtain a typical load curve.

[0007] This scheme proposes a power load clustering method and system, which can process load time series data with asynchronous collection time, unequal time length and noise. However, this method can only be applied to the processing of a single power load type. The present invention can identify and distinguish load types with different energy consumption modes.

[0008] The application number is CN202211127930.4, and the name is a load clustering method based on feature extraction and improved K-means algorithm. The present invention discloses a load clustering method based on feature extraction and improved K-means algorithm, which relates to the technical field of load clustering of power systems, and specifically relates to a load clustering method based on feature extraction and improved K-means algorithm. By inputting power load data, calculating load characteristic indicators, using singular value decomposition method to reduce the dimension of load data, using the load data after dimension reduction, and using the improved K-means algorithm considering density to perform load clustering. The present invention solves the problems of unstable clustering results, the risk of losing important features of the original load, and high computational complexity in the current load clustering methods, and provides an effective quantitative reference for the next step of load forecasting, energy storage capacity configuration, demand-side management, etc.

[0009] This solution proposes a load clustering method based on feature extraction and improved K-means algorithm, which solves the problems of unstable clustering results, the risk of losing important features of the original load, and high computational complexity in the current load clustering method. However, the use of SEE (and variance) as the evaluation index of the improved K-means algorithm does not directly involve the relative position relationship or distance between different categories in the clustering results, and cannot directly reflect the relative relationship between and within the clustering results, and cannot effectively distinguish multivariate loads. The present invention uses the CH (clustering effect evaluation) index as an evaluation index, which can reflect the distance between the difference matrices within and between different categories, thereby effectively distinguishing multivariate loads. Summary of the invention

[0010] In order to solve the problems that the traditional load clustering methods in the prior art have unstable clustering results, the risk of losing important features of the original load, and high computational complexity, and are difficult to meet the needs of rapid extraction of large-scale data; there are many factors of cold, hot and electric loads on the demand side, and the existing methods are difficult to effectively distinguish the multi-loads, the present invention proposes a regional demand side multi-load clustering method, which uses the CH-K-means clustering method to perform clustering analysis of typical days selected throughout the year for multi-loads, including the following steps:

[0011] Collect historical data of multi-loads in typical seasons of the park and temperature data and holiday information at corresponding time nodes;

[0012] Calculating the multivariate load historical mean according to the multivariate load historical data, judging the authenticity of the data according to the 3σ principle combined with a set threshold, preprocessing the multivariate load historical data, and obtaining the preprocessed multivariate load historical data;

[0013] The CH-K-means clustering method is used to cluster the multivariate load curves of the preprocessed multivariate load historical data using the clustering effect evaluation index as the best clustering result evaluation index to obtain the clustering results.

[0014] An analysis is performed based on the clustering results to form a multivariate load distribution.

[0015] Furthermore, the multi-load historical data includes historical data of cooling load, heating load and electric load.

[0016] Furthermore, the method for clustering multi-loads on the regional demand side includes: calculating the multi-load historical mean value according to the multi-load historical data as shown in the following formula:

[0017]

[0018] In the formula, X n,i is the multivariate load historical data; N is the total number of historical days; is the historical mean of multivariate load; i is the load type; n is the number of days.

[0019] Further, the determining of data authenticity includes: calculating a standard deviation σ before determining the data authenticity, as shown in the following formula:

[0020]

[0021] In the formula, X n,i is the multivariate load historical data; N is the total number of historical days; is the historical mean of multivariate load; i is the load type; n is the number of days; σ is the standard deviation.

[0022] The data authenticity is judged according to the 3σ principle combined with the set threshold as shown in the following formula:

[0023]

[0024] Where, X n,i is the historical data of multivariate load; is the historical mean of multivariate load; σ i is the standard deviation of load of type i; ε is the threshold; i is the load type; n is the number of days;

[0025] Furthermore, the preprocessing of the multivariate load historical data includes: n,i , use the repair formula to correct the weights, and n,i Perform noise detection and data repair to obtain repaired data.

[0026] Furthermore, the repair formula is as follows:

[0027]

[0028] Where, α+β+γ=1; is the load correction data of the nth day for the i-type load; N is the total number of historical days; X n,i is the historical data of multivariate load; X n,i-1 and X n,i+1 They are the two nearest transverse multi-load points; Yes X n,i 4 similar daily multivariate load points nearby; is the historical mean of multivariate load; i is the load type; n is the number of days.

[0029] Furthermore, the preprocessing of the multivariate load historical data includes: dimensionlessly processing the repaired multivariate load historical data and temperature data into a data set matrix X∈(0,1) and normalizing them, and setting the data set matrix to X, as shown in the following formula:

[0030]

[0031] Where u is the total number of training samples; v is the number of input features.

[0032] Furthermore, the normalization process includes: normalizing each column of the dimensionless processed data set matrix X, as shown in the following formula:

[0033]

[0034] Where, X Ris the normalized data set matrix; u is the total number of training samples; v is the number of input features; x pq is a class of characteristic numbers; is the number of features after normalization; p is the input sequence of training samples; q is the input feature sequence; 1≤p≤u; 1≤q≤v; is the minimum value of the number of features of a class; is the range of a class of characteristic numbers.

[0035] Furthermore, the CH-K-means clustering method is used to cluster the multivariate load curves of the preprocessed multivariate load historical data using the clustering effect evaluation index as the evaluation index of the best clustering result, and the clustering results are obtained, including:

[0036] Determine the search range for the optimal number of clusters;

[0037] The clustering effect evaluation index is selected as the evaluation index of the best clustering result;

[0038] For loads with the same mark, take their average hourly value to obtain the average load curve, which is recorded as the typical daily load curve.

[0039] Furthermore, the determining of the optimal cluster number search range includes:

[0040] Set the number of clusters to k well , k well ∈[2,k max ],Regulation Where n1 is the number of samples.

[0041] Furthermore, the evaluation index of the best clustering result includes:

[0042] The evaluation index of the best clustering result is marked as CH(k), which is calculated as follows:

[0043]

[0044] In the formula, CH(k) is the closeness between data; k is the number of clusters; n1 is the number of samples; TrN(k) is the distance of the deviation matrix within the class, and TrJ(k) is the distance of the deviation matrix between classes.

[0045] Furthermore, the typical daily load curve is calculated as follows:

[0046]

[0047] In the formula, is the typical daily load curve; n2 is the number of similar clustering days; t is the clustering category; is the load value at a certain moment; i1 is the number of days, and j1 is the number of hours.

[0048] Furthermore, the analysis is performed based on the clustering results to form a multivariate load distribution, including: performing analysis based on the clustering results, comparing the typical daily load curve with the known classic load curve, judging whether it is a classic load pattern based on the similarity between the curves, and if so, forming a multivariate load distribution; if not, returning to the step of judging the data authenticity based on the 3σ principle combined with the set threshold.

[0049] The present invention provides a regional demand side multi-load clustering system, the system comprising:

[0050] Processing module: used to obtain the historical data of multivariate loads in a typical season of the park and the temperature data and holiday information at the corresponding time nodes, and to perform data processing based on the historical data of multivariate loads to obtain a data set matrix for clustering;

[0051] Clustering module: used for selecting clustering effect evaluation index as the result evaluation index to perform multivariate load curve clustering on the preprocessed multivariate load historical data;

[0052] Result module: used to obtain the typical daily load curve of clustering results.

[0053] Wherein, the processing module includes:

[0054] Acquisition unit: Acquisition of historical data of multi-loads in typical seasons of the park and temperature data and holiday information at corresponding time nodes;

[0055] Screening unit: used for judging the authenticity of the multivariate load historical data by adopting the 3σ principle;

[0056] Correction unit: used to detect noise points and repair data using a repair formula for distorted data;

[0057] Conversion unit: used for normalizing the data repaired by the repair formula to obtain pre-processed multivariate load historical data.

[0058] In another aspect, the present application further provides an electronic device, comprising: at least one processor and a memory, wherein the memory and the processor are connected via a bus;

[0059] The memory is used to store one or more programs;

[0060] When the one or more programs are executed by the at least one processor, a regional demand side multi-load clustering method as described above is implemented.

[0061] On the other hand, the present application also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed, a regional demand side multi-load clustering method as described above is implemented.

[0062] Compared with the prior art, the present invention has the following beneficial effects:

[0063] The present invention discloses a method and system for clustering multivariate loads on the regional demand side, including: collecting multivariate load historical data of a typical season in a park area and temperature data and holiday information at corresponding time nodes; calculating the multivariate load historical mean according to the multivariate load historical data, and judging the authenticity of the data according to the 3σ principle combined with a set threshold, preprocessing the multivariate load historical data to obtain processed multivariate load historical data; using the CH-K-means clustering method, using the clustering effect evaluation index as the evaluation index of the best clustering result to cluster the multivariate load curves of the preprocessed multivariate load historical data to obtain a clustering result; analyzing according to the clustering result to form a multivariate load distribution. The present invention uses the CH-K-means clustering method to perform clustering analysis of typical days selected throughout the year for multivariate loads, meeting the demand for rapid extraction of large-scale data; mining load data characteristic indicators, using the clustering effect evaluation index as the best clustering result to cluster the multivariate load curves, solving the problem that there are many factors on the demand side cold, heat, and electricity loads that make it difficult to effectively distinguish multivariate loads. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a schematic diagram of the regional demand side multi-load clustering method of the present invention;

[0065] Figure 2 It is a flow chart of the regional demand side multi-load clustering method of the present invention;

[0066] Figure 3 It is a schematic diagram of the structure of the electronic device of the present invention. DETAILED DESCRIPTION

[0067] The present invention provides a regional demand side multivariate load clustering method, which adopts the CH-K-means clustering method to perform typical day clustering analysis of multivariate loads throughout the year to meet the demand for rapid extraction of scale data.

[0068] The present invention aims to accurately classify and optimize the aggregation according to different types of load characteristics to achieve more efficient energy utilization and load scheduling. By clustering the load data, the method can identify and distinguish the load types of different energy consumption modes, and aggregate them to form a multivariate load distribution, providing detailed basic data for intelligent load management.

[0069] In order to better understand the present invention, the content of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0070] Example 1

[0071] The present invention proposes a regional demand side multi-load clustering method, such as Figure 1 As shown, the specific implementation method includes:

[0072] Step 1: Collect the historical data of multi-loads in the typical season of the park and the temperature data and holiday information at the corresponding time nodes;

[0073] Step 2: Calculate the multivariate load historical mean according to the multivariate load historical data, and judge the data authenticity according to the 3σ principle combined with the set threshold, pre-process the multivariate load historical data, and obtain the pre-processed multivariate load historical data;

[0074] Step 3: Using the CH-K-means clustering method, the clustering effect evaluation index is used as the evaluation index of the best clustering result to cluster the multivariate load curves of the preprocessed multivariate load historical data to obtain the clustering results;

[0075] Step 4: Analyze the clustering results to form a multivariate load distribution.

[0076] In order to better describe the present invention, the following specific embodiments are combined with Figure 2 The present invention is further introduced.

[0077] Step 1: Collect the historical data of multi-loads in the typical season of the park and the temperature data and holiday information at the corresponding time nodes;

[0078] The historical data of the multi-loads include historical data of cooling load, heating load and electric load.

[0079] Step 2: Calculate the multivariate load historical mean according to the multivariate load historical data, judge the data authenticity according to the 3σ principle combined with the set threshold, pre-process the multivariate load historical data, and obtain processed multivariate load historical data.

[0080] The preprocessing of the multivariate load historical data includes: performing noise point detection and data repair on the multivariate load historical data, and performing normalization processing on the repaired data.

[0081] Step 2 specifically includes:

[0082] 1. Calculate the historical mean of multivariate load

[0083]

[0084] 2. Calculate the standard deviation σ:

[0085]

[0086] In formulas (1) and (2), X n,i is the historical data of multivariate load; is the historical mean of multivariate load; N is the total number of historical days; σ is the standard deviation; i is the load type; n is the number of days.

[0087] 3. According to the 3σ principle and the set threshold, the authenticity of the data is judged by formula (3):

[0088]

[0089] In formula (3), X n,i is the historical data of multivariate load; is the historical mean of multivariate load; σ i is the standard deviation of load of type i; ε is the threshold; i is the load type; n is the number of days;

[0090] The threshold ε ranges from 1 to 1.5.

[0091] 4. Satisfy formula (3) to determine the multi-element load historical data X n,i Distortion, the weight correction is performed using formula (4), and the multivariate load historical data X n,i Perform noise detection and data repair:

[0092]

[0093] In formula (4), α+β+γ=1; is the load correction data of the nth day for the i-type load; N is the total number of historical days; X n,i is the historical data of multivariate load; X n,i-1 and X n,i+1 They are the two nearest transverse multi-load points; Yes X n,i 4 similar daily multivariate load points nearby; is the historical mean of multivariate load; i is the load type; n is the number of days.

[0094] 5. The repaired multivariate load historical data and temperature data are dimensionlessly processed into a data set matrix X∈(0,1) and normalized. The data set matrix is ​​set to X, as shown in the following formula:

[0095]

[0096] In formula (5), u is the total number of training samples; v is the number of input features.

[0097] 6. Normalize each column of the processed data set matrix X, as shown in the following formula:

[0098]

[0099]

[0100] In formulas (6) and (7), X R is the normalized data set matrix; u is the total number of training samples; v is the number of input features; x pq is a class of characteristic numbers; is the number of features after normalization; p is the input sequence of training samples; q is the input feature sequence; 1≤p≤u; 1≤q≤v; is the minimum value of the number of features of a class; is the range of a class of characteristic numbers.

[0101] Step 3: Use the CH-K-means clustering method and the clustering effect evaluation index as the evaluation index of the best clustering result to cluster the multivariate load curves of the preprocessed multivariate load historical data to obtain the clustering results; the specific process includes the following:

[0102] Determine the search range for the optimal number of clusters;

[0103] The clustering effect evaluation index is selected as the evaluation index of the best clustering result;

[0104] For loads with the same mark, take their average hourly value to obtain the average load curve, which is recorded as the typical daily load curve.

[0105] Step 3 specifically includes:

[0106] 1. Determine the optimal cluster number search range and set the cluster number to k well , k well ∈[2,k max ],Regulation Where n1 is the number of samples;

[0107] 2. The CH (Calinski-Harabasz, clustering effect evaluation) index is selected as the evaluation index of the best clustering result, denoted as CH(k), and the calculation is as follows:

[0108]

[0109] In formula (8), CH(k) is the closeness between data; k is the number of clusters; n1 is the number of samples; TrN(k) is the distance of the intra-class deviation matrix, and TrJ(k) is the distance of the inter-class deviation matrix.

[0110] 3. Take the average of the loads with the same mark hourly to obtain the average load curve, which is recorded as the typical daily load curve and calculated as follows:

[0111]

[0112] In formula (9), is the typical daily load curve; n2 is the number of similar clustering days; t is the clustering category; is the load value at a certain moment; i1 is the number of days, and j1 is the number of hours.

[0113] Step 4: Analyzing the clustering results to form a multivariate load distribution, specifically includes:

[0114] According to the clustering results, the typical daily load curve is compared with the known classic load curve. The similarity between the curves is used to determine whether it is a classic load pattern. If so, a multivariate load distribution is formed; if not, return to step 2.

[0115] Embodiment 2:

[0116] The present invention based on the same inventive concept also provides a regional demand side multi-load clustering system, the system comprising:

[0117] Processing module: used to obtain the multi-load historical data of the park in a typical season and the temperature data and holiday information at the corresponding time node and pre-process the multi-load historical data to obtain the pre-processed multi-load historical data;

[0118] Clustering module: used for selecting clustering effect evaluation index as the result evaluation index to perform multivariate load curve clustering on the preprocessed multivariate load historical data;

[0119] Result module: used to obtain the typical daily load curve of clustering results.

[0120] Wherein, the processing module includes:

[0121] Acquisition unit: Acquisition of historical data of multi-loads in typical seasons of the park and temperature data and holiday information at corresponding time nodes;

[0122] Screening unit: used for judging the authenticity of the multivariate load historical data by adopting the 3σ principle;

[0123] Correction unit: used to detect noise points and repair data using a repair formula for distorted data;

[0124] Normalization unit: used to normalize the data repaired by the repair formula to obtain pre-processed multivariate load historical data.

[0125] Wherein, the clustering module includes:

[0126] Limit unit: used to determine the search range for the optimal number of clusters;

[0127] Optimization unit: used to use CH (Calinski-Harabasz, best clustering result) as the evaluation index;

[0128] Calculation unit: used to obtain the hourly average of loads of the same type.

[0129] Wherein, the result module includes:

[0130] Result unit: used to obtain the typical daily load curve based on the data calculated by the calculation unit.

[0131] Example 3

[0132] like Figure 3 As shown, the present invention also provides an electronic device, which may be a computer device, a single-chip device, an intelligent mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, the processor, and the transceiver component are connected via a bus; the memory may be used to store an execution program, and an exemplary execution program may include instructions; the processor is used to execute the instructions stored in the memory. The memory may also be used to store data, which may be called and / or modified when the instructions are executed.

[0133] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), or application-specific integrated circuits.

[0134] (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, and are suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the storage medium to implement the corresponding method flow or corresponding function, so as to implement the steps of a regional demand side multi-load clustering method in the above embodiment.

[0135] Example 4

[0136] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It can be understood that the storage medium here can include both built-in storage media in electronic devices and extended storage media supported by electronic devices. The storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk storage. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a regional demand side multi-load clustering method in the above embodiment.

[0137] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0138] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0139] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0140] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0141] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims

1. A regional demand side multi-load clustering method, characterized in that: The method comprises: Collect historical data of multi-loads in typical seasons of the park and temperature data and holiday information at corresponding time nodes; Calculating the multivariate load historical mean according to the multivariate load historical data, judging the authenticity of the data according to the 3σ principle combined with a set threshold, preprocessing the multivariate load historical data, and obtaining the preprocessed multivariate load historical data; The CH-K-means clustering method is adopted, and the clustering effect evaluation index is used as the evaluation index of the best clustering result to perform multivariate load curve clustering on the preprocessed multivariate load historical data to obtain a clustering result; An analysis is performed based on the clustering results to form a multivariate load distribution.

2. A regional demand side multi-load clustering method according to claim 1, characterized in that: The calculation of the multi-element load historical mean value based on the multi-element load historical data is shown in the following formula: In the formula, X n,i is the multivariate load historical data; N is the total number of historical days; is the historical mean of multivariate load; i is the load type; n is the number of days.

3. A regional demand side multi-load clustering method according to claim 1, characterized in that: Determine the authenticity of the data according to the following formula: Where, X n,i is the multivariate load historical data; is the historical mean of multivariate load; σ i is the standard deviation of load type i; ε is the threshold; i is the load type; n is the number of days.

4. The method according to claim 3, characterized in that The preprocessing of the multivariate load historical data includes: n,i , use the repair formula to correct the weights, and n,i Perform noise detection and data repair to obtain repaired data.

5. The method according to claim 4, characterized in that The repair formula is as follows: Where, α+β+γ=1; is the load correction data of the nth day for the i-type load; N is the total number of historical days; X n,i is the historical data of multivariate load; X n,i-1 and X n,i+1 are the two nearest transverse multi-load points; Yes X n,i 4 similar daily multivariate load points nearby; is the historical mean of multivariate load; i is the load type; n is the number of days.

6. A regional demand side multi-load clustering method according to claim 4, characterized in that: The preprocessing of the multivariate load historical data includes: dimensionlessly processing the repaired multivariate load historical data and temperature data into a data set matrix X∈(0,1), and performing normalization processing to obtain the preprocessed multivariate load historical data.

7. A regional demand side multi-load clustering method according to claim 1, characterized in that: The CH-K-means clustering method is adopted, and the clustering effect evaluation index is used as the evaluation index of the best clustering result to cluster the multivariate load curves of the pre-processed multivariate load historical data, and the clustering results obtained include: Determine the search range for the optimal number of clusters; The clustering effect evaluation index is selected as the evaluation index of the best clustering result; For loads with the same mark, take their average hourly value to obtain the average load curve, which is recorded as the typical daily load curve.

8. The method according to claim 7, characterized in that The evaluation index of the best clustering result is calculated as follows: In the formula, CH(k) is the closeness between data; k is the number of clusters; n1 is the number of samples; TrN(k) is the distance of the deviation matrix within the class, and TrJ(k) is the distance of the deviation matrix between classes.

9. The method according to claim 7, characterized in that: The typical daily load curve is calculated as follows: In the formula, is the typical daily load curve; n2 is the number of similar clustering days; t is the cluster category; is the load value at a certain moment; i1 is the number of days, and j1 is the number of hours.

10. A regional demand side multi-load clustering system, characterized in that: The system comprises: Processing module: used to obtain the multi-load historical data of the park in a typical season and the temperature data and holiday information at the corresponding time node and pre-process the multi-load historical data to obtain the pre-processed multi-load historical data; Clustering module: used for selecting clustering effect evaluation index as the result evaluation index to perform multivariate load curve clustering on the preprocessed multivariate load historical data; Result module: used to obtain the typical daily load curve of clustering results.

11. The system according to claim 10, characterized in that The processing module comprises: Acquisition unit: Acquisition of historical data of multi-loads in typical seasons of the park and temperature data and holiday information at corresponding time nodes; Screening unit: used for judging the authenticity of the multivariate load historical data by adopting the 3σ principle; Correction unit: used to detect noise points and repair data using a repair formula for distorted data; Conversion unit: used for normalizing the data repaired by the repair formula to obtain pre-processed multivariate load historical data.

12. An electronic device, characterized in that: include: at least one processor and memory; The memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, a regional demand side multi-element load clustering method as described in any one of claims 1 to 9 is implemented.

13. A readable storage medium, characterized in that: An execution program is stored thereon, and when the execution program is executed, a regional demand side multi-load clustering method as described in any one of claims 1 to 9 is implemented.

Images