A daily reactive load curve clustering method and system based on feature index dimension reduction

Abstract

The application provides a daily reactive load curve clustering method and system based on feature index dimension reduction, and uses load rate, highest utilization hour rate, daily peak-valley difference rate, peak period load rate, maximum load occurrence time, flat period load rate, valley period load rate and minimum load occurrence time as daily reactive load feature indexes; daily reactive load data is reduced based on the daily reactive load feature indexes, a daily reactive load feature index data reduced matrix is obtained, and an expert weighting method is used to assign weights to the daily reactive load feature indexes; and a clustering calculation is performed based on a k-means clustering algorithm and a daily reactive load curve clustering result is output. The established daily reactive load feature index system is applied to the daily reactive load clustering calculation, the calculation amount is effectively reduced, the problem that the clustering effect is not ideal under the high-dimensional data condition is avoided, the daily reactive load can be accurately clustered, and the daily reactive load clustering has a guiding effect on formulating a load dispatching plan and operation control.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of data processing, and particularly relates to a daily reactive load curve clustering method and system based on feature index dimension reduction. BACKGROUND

[0002] The statements in this section merely provide background information related to the present application and do not necessarily constitute prior art.

[0003] In the research of power system load characteristics, an important research is to obtain the typical daily load curve. The typical daily load curve can be used to analyze the characteristics of regional load power consumption, obtain the typical daily maximum (minimum) load and average load, and has important significance for formulating load scheduling plan and operation control. However, the present application finds that the current research on daily load curve mainly focuses on active load, and the research on reactive load curve is less, and there are the following problems:

[0004] For the typical daily active load curve, there are three conventional selection methods, that is, 1) selecting the daily load curve closest to the monthly average load rate and without distortion as the typical daily load curve of the month; 2) selecting the daily load curve with the maximum load of the month as the daily load curve of the month; and 3) selecting a fixed working day of the month as the typical daily load curve of the month. However, the above methods have certain limitations and cannot fully reflect the load characteristics in a period.

[0005] For example, when the daily load curve closest to the monthly average load rate and without distortion is selected as the typical daily load curve of the month, on the one hand, only one daily load curve (even the closest one) is selected as the typical representative of the entire month, which cannot fully reflect the diversity and variability of the load in the month. Because in actual life, even if the monthly average load rate is close, the load curves of different dates may still have significant differences, such as holidays, special weather, changes in economic activities, etc., which may affect the load; on the other hand, if the selected daily load curve does not contain or contains few extreme load conditions (such as the highest load day, the lowest load day, etc.), the reference value of the typical daily load curve in dealing with extreme load will be limited.

[0006] In addition, the k-means clustering algorithm is a powerful means of organizing and managing disordered data, and is effectively applied to load curve clustering calculation. However, the reactive load curve is multi-dimensional data, and the clustering effect is often not ideal when the conventional clustering method is directly applied, so appropriate means need to be taken to reduce the dimension of the reactive load curve. SUMMARY

[0007] In order to overcome the above-mentioned deficiencies of the prior art, the present application provides a daily reactive load curve clustering method and system based on feature index dimension reduction, which can accurately cluster the daily reactive load curve.

[0008] To achieve the above object, one or more embodiments of the present application provide the following technical solutions:

[0009] The first aspect of the present application provides a daily reactive load curve clustering method based on feature index dimension reduction, comprising:

[0010] Obtaining the annual reactive load raw data of the transformer substation;

[0011] Taking at least eight indexes of load rate, highest utilization hour rate, daily peak-valley difference rate, peak period load rate, maximum load occurrence time, flat period load rate, valley period load rate and minimum load occurrence time as the daily reactive load feature indexes;

[0012] Based on the daily reactive load feature indexes, the daily reactive load data is dimensionally reduced to obtain a daily reactive load feature index data dimension reduction matrix, and an expert weighting method is used to assign weights to the daily reactive load feature indexes;

[0013] Based on the k-means clustering algorithm, the dimensionally reduced daily reactive load data is clustered and calculated, and when the judgment condition for ending clustering is met, the clustering is stopped and the daily reactive load curve clustering result is output.

[0014] The second aspect of the present application provides a daily reactive load curve clustering system based on feature index dimension reduction, comprising:

[0015] The data acquisition module is configured to obtain the annual reactive load raw data of the transformer substation.

[0016] The feature index selection module is configured to take at least eight indexes of load rate, highest utilization hour rate, daily peak-valley difference rate, peak period load rate, maximum load occurrence time, flat period load rate, valley period load rate and minimum load occurrence time as the daily reactive load feature indexes.

[0017] The data processing module is configured to dimensionally reduce the daily reactive load data based on the daily reactive load feature indexes and obtain a daily reactive load feature index data dimension reduction matrix, and to assign weights to the daily reactive load feature indexes using an expert weighting method.

[0018] The clustering module is configured to cluster and calculate the dimensionally reduced daily reactive load data based on the k-means clustering algorithm, and to stop clustering and output the daily reactive load curve clustering result when the judgment condition for ending clustering is met.

[0019] The third aspect of the present application provides a computer readable storage medium, which stores a program, and the program is executed by a processor to implement the steps of the feature index dimension reduction based daily reactive load curve clustering method according to the first aspect of the present application.

[0020] The fourth aspect of the present application provides an electronic device, which comprises a memory, a processor and a program stored in the memory and executable on the processor, and the processor executes the program to implement the steps of the feature index dimension reduction based daily reactive load curve clustering method according to the first aspect of the present application.

[0021] The above one or more technical solutions have the following beneficial effects.

[0022] (1) The eight indexes that can best represent the load characteristics of the daily reactive load data are used as the basis to reduce the dimension of the reactive load data, so that the multi-dimensional characteristics of the reactive load curve are considered, and the problem of unsatisfactory clustering effect caused by direct clustering is avoided. Meanwhile, on the basis of dimension reduction, the simple, high-precision and fast k-means clustering algorithm is selected for the final clustering operation, so that the clustering effect is good and the efficiency is higher, and the daily reactive load curve can be accurately clustered, which has a guiding effect on the formulation of load scheduling plan and operation control.

[0023] (2) The eight load characteristic indexes are applied to the daily reactive load clustering calculation, so that the calculation amount is effectively reduced, and the problem that only the load characteristics of the active load curve are analyzed in the prior art, and the analysis of the related load characteristics of the reactive load is ignored is solved.

[0024] The advantages of the additional aspects of the present application will be partially given in the following description, partially become obvious from the following description, or be understood through the practice of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0025] The accompanying drawings, which form a part of the present application, are used to provide further understanding of the present application, and the illustrative embodiments of the present application and their description are used to explain the present application, and do not constitute improper limitations on the present application.

[0026] Figure 1 The flowchart of the feature index dimension reduction based daily reactive load curve clustering method in example one.

[0027] Figure 2 The flowchart of the k-means clustering algorithm in example one.

[0028] Figure 3 The module schematic diagram of the feature index dimension reduction based daily reactive load curve clustering system in example two. DETAILED DESCRIPTION

[0029] It should be noted that the following detailed description is exemplary in nature and is intended to provide further description of the application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0030] It should be noted that the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application.

[0031] In the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.

[0032] The present application provides a day reactive load curve clustering method and system based on feature index dimension reduction, and the overall idea is:

[0033] Obtain the annual reactive load raw data of the substation; at least eight indexes of load rate, highest utilization hour rate, daily peak valley difference rate, peak period load rate, maximum load occurrence time, flat period load rate, valley period load rate and minimum load occurrence time are used as day reactive load feature indexes. Based on the day reactive load feature indexes, the day reactive load data is reduced and the day reactive load feature index data reduction matrix is obtained, and the expert weighting method is used to assign weights to the day reactive load feature indexes. Based on the k-means clustering algorithm, the day reactive load data after dimension reduction is clustered and calculated, and when the judgment condition of clustering end is met, the clustering is stopped and the day reactive load curve clustering result is output. Through the above means, the day reactive load curve clustering can be accurately performed.

[0034] In order to facilitate the understanding of the technical scheme of the present application, the present application will be further described below in combination with embodiments.

[0035] Embodiment one

[0036] As shown in Figure 1 The present embodiment provides a day reactive load curve clustering method based on feature index dimension reduction, which can be realized by the following process:

[0037] Step S1, obtaining the annual reactive load raw data of the substation.

[0038] In the actual acquisition of the annual reactive load raw data of the substation, the annual reactive load data of the substation is collected, specifically, the sampling interval is set to 15 minutes, and 96 reactive load data will be collected every day, and then the annual reactive load raw data matrix is obtained, that is:

[0039]

[0040] Among them, X i represents the reactive load data set of the i-th day; xij Qi,j represents the reactive power load data at the jth time point of the ith day; a represents the total number of days in the year in which the reactive power load data is located, b represents the number of reactive power load data collected in a day, thus a = 365, b = 96.

[0041] Step S2, selecting a daily reactive power load characteristic index.

[0042] The load characteristics of a power system refer to the characteristics and properties of the power load, which can be embodied by relevant characteristic indexes. At present, there is no unified standard for the load characteristics of a power system internationally, and generally, 15 indexes such as load rate, peak-valley difference, and average load are used to analyze the characteristics and properties of various power loads.

[0043] The reactive power load curve and the active power load curve have certain similarities, and both have characteristics such as load rate, peak-valley difference rate, and peak period load rate, so in this embodiment, the load rate, the highest utilization hour rate, the daily peak-valley difference rate, the peak period load rate, the maximum load occurrence time, the flat period load rate, the valley period load rate, and the minimum load occurrence time are taken as the daily reactive power load characteristic indexes.

[0044] The daily reactive power load characteristic indexes are characteristic indexes in different time periods, and are specifically shown in Table 1:

[0045] Table 1 Daily reactive power load characteristic indexes

[0046]

[0047] Wherein, Q represents the reactive power load; t represents the time; the subscripts sum, av, max, and min represent the total amount, the average value, the maximum value, and the minimum value respectively; the superscripts peak, val, and sh represent the peak period, the valley period, and the flat period respectively.

[0048] Based on the obtained substation annual reactive power load raw data, the average value of the reactive power load in each time period in the whole year is calculated, and the time periods in the whole day are divided into peak period time periods, flat period time periods, and valley period time periods according to the average value.

[0049] Specifically, in this embodiment, the time periods of 08:00-11:00 and 18:00-21:00 in a day are taken as the peak period time periods, the time periods of 06:00-08:00, 11:00-18:00, and 21:00-22:00 in a day are taken as the flat period time periods, and the time periods of 22:00-24:00 and 00:00-06:00 in a day are taken as the valley period time periods.

[0050] From Table 1, the load rate, the highest utilization hour rate and the daily peak valley difference rate are the daily reactive load characteristic indexes in the whole day period, the peak period load rate and the maximum load occurrence time are the daily reactive load characteristic indexes in the peak period, the flat period load rate is the daily reactive load characteristic index in the flat period, and the valley period load rate and the minimum load occurrence time are the daily reactive load characteristic indexes in the valley period.

[0051] In step S3, the daily reactive load data is reduced in dimension based on the daily reactive load characteristic indexes, and a daily reactive load characteristic index data reduced matrix is obtained, and an expert weighting method is used to assign weights to the daily reactive load characteristic indexes.

[0052] Specifically, the daily reactive load characteristic index data reduced matrix is represented as:

[0053]

[0054] Wherein, Y represents the daily reactive load characteristic index data reduced matrix, a represents the total number of days in the year of the annual reactive load data, Y a represents the reactive load characteristic index reduced data set of the a-th day, c represents the number of daily reactive load characteristic indexes, i.e. c = 8, y ac represents the c-th daily reactive load characteristic index value of the a-th day.

[0055] From the daily reactive load characteristic index data reduced matrix, it can be seen that the daily reactive load data is characterized by 96 sampling data per day in a year, which is changed to 8 daily reactive load characteristic indexes per day in a year, i.e. in a year of 365 days, the daily reactive load characteristic index data reduced matrix reduces the original data of the annual reactive load of the substation from a high-dimensional data matrix of 365*96 to a data matrix of 365*8, greatly reducing the dimension of the original data of the annual reactive load of the substation.

[0056] The expert weighting method is used to assign weights to the daily reactive load characteristic indexes, and the obtained daily reactive load characteristic index weight matrix is:

[0057]

[0058] Wherein, W represents the daily reactive load characteristic index weight matrix, W m represents the reactive load characteristic index reduced weight data set of the m-th expert, d represents the number of invited experts, c represents the number of daily reactive load characteristic indexes, w ml represents the weight assigned by the m-th expert to the l-th daily reactive load characteristic index, which satisfies

[0059] The characteristic index weight mean value of all experts is calculated based on the obtained weight matrix of daily reactive load characteristic indexes, and the characteristic index weight mean value is composed of the weight mean value given by all experts to each daily reactive load characteristic index, that is:

[0060]

[0061] wherein, represents the weight mean value given by all experts to the lth daily reactive load characteristic index.

[0062] Finally, the characteristic index weight mean value matrix of all experts is calculated, that is:

[0063]

[0064] wherein, W ave represents the characteristic index weight mean value matrix of all experts, and c represents the number of daily reactive load characteristic indexes.

[0065] Step S4: based on the k-means clustering algorithm, the dimension-reduced daily reactive load data is clustered and calculated, and when the judgment condition of clustering end is met, the clustering is stopped and the daily reactive load curve clustering result is output.

[0066] The number of clusters k (k is an integer, The Euclidean distance is used as an index for measuring the similarity between data objects.

[0067] As Figure 2 shown, when the k-means clustering algorithm is used for clustering calculation, the following steps are performed:

[0068] 1) Let k = 1, and use the reactive load characteristic index dimension-reduced data Y, the expert characteristic index weight W ave as input data;

[0069] 2) Perform k = k + 1, and then randomly generate k initial cluster centers, that is:

[0070]

[0071] wherein, Z n represents the nth cluster center; z nl represents the lth characteristic index value of the nth cluster center.

[0072] 3) The weighted Euclidean distance between each data object and each cluster center Z n is calculated by the following formula, that is:

[0073]

[0074] 4) According to the minimum distance principle, reclassify each data object;

[0075] 5) Calculate the mean of each class, and update the data center of each class;

[0076] 6) Calculate the sum of squared errors SSE of the entire data set, that is:

[0077]

[0078] Where a represents the total number of days in the year of the annual reactive load data, k represents the number of clusters, dis(Y i Z n ) represents the weighted Euclidean distance between the i-th day of the day reactive load feature index dimension reduction data set Y i and the n-th cluster center Z n .

[0079] 7) Determine whether SSE changes? Yes, return to step 3; No, proceed to the next step.

[0080] 8) Determine ? Yes, return to step 2; No, proceed to the next step.

[0081] 9) Generate the SSE curve with k value, and take the "elbow point" in the change curve of the sum of squared errors SSE with k value as the optimal cluster number and output the clustering result.

[0082] Example Two

[0083] The embodiment discloses a day reactive load curve clustering system based on feature index dimension reduction.

[0084] As shown in Figure 3 , a day reactive load curve clustering system based on feature index dimension reduction comprises:

[0085] A data acquisition module configured to obtain substation annual reactive load original data.

[0086] A feature index selection module configured to take load rate, highest utilization hour rate, day peak valley difference rate, peak period load rate, maximum load occurrence time, flat period load rate, valley period load rate, and minimum load occurrence time as day reactive load feature indexes.

[0087] A data processing module configured to perform day reactive load data dimension reduction based on day reactive load feature indexes and obtain day reactive load feature index data dimension reduction matrix, and to assign weights to day reactive load feature indexes using an expert weighting method.

[0088] The clustering module is configured to perform clustering calculation on the dimension-reduced daily reactive load data based on a k-means clustering algorithm, and stop clustering and output a daily reactive load curve clustering result when a judgment condition for ending clustering is met.

[0089] Embodiment three

[0090] An object of the embodiment is to provide a computer-readable storage medium.

[0091] A computer-readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the steps in the daily reactive load curve clustering method based on dimension reduction of feature indicators according to Embodiment 1 of the present disclosure.

[0092] Embodiment four

[0093] An object of the embodiment is to provide an electronic device.

[0094] An electronic device, comprising a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor implements the steps in the daily reactive load curve clustering method based on dimension reduction of feature indicators according to Embodiment 1 of the present disclosure when executing the program.

[0095] The steps and methods involved in the above embodiments two, three, and four correspond to Embodiment 1, and the specific implementation can be referred to the relevant description part of Embodiment 1. The term “computer-readable storage medium” should be understood as including a single medium or multiple media of one or more instruction sets; and should also be understood as including any medium capable of storing, encoding, or carrying instruction sets for execution by a processor and causing the processor to perform any method in the present disclosure.

[0096] Those skilled in the art should understand that each module or step of the present disclosure described above can be implemented by a general computer device, and alternatively, they can be implemented by program codes executable by a computing device, so that they can be stored in a storage device for execution by a computing device, or they can be respectively manufactured into individual integrated circuit modules, or a plurality of modules or steps among them can be manufactured into a single integrated circuit module. The present disclosure is not limited to any specific combination of hardware and software.

[0097] The above describes the specific embodiments of the present disclosure in conjunction with the accompanying drawings, but is not a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications or variations made by those skilled in the art on the basis of the technical solutions of the present disclosure without inventive labor are still within the protection scope of the present disclosure.

Claims

1. A method for clustering daily reactive load curves based on feature indicators dimension reduction, characterized in that, The method comprises the following steps: obtaining annual reactive power load raw data of a substation; at least eight indexes of load rate, maximum utilization hour rate, daily peak-valley difference rate, peak period load rate, maximum load occurrence time, flat period load rate, valley period load rate and minimum load occurrence time are taken as daily reactive power load characteristic indexes; dimensionality reduction is performed on daily reactive power load data based on daily reactive power load characteristic indexes to obtain a daily reactive power load characteristic index data reduction matrix, and expert weighting method is used to assign weights to daily reactive power load characteristic indexes; when a judgment condition for ending clustering is met, clustering is stopped and a daily reactive power load curve clustering result is outputted.

2. The method of claim 1, wherein the method is based on feature index dimension reduction of daily reactive load curve clustering. Based on the obtained annual reactive power load raw data of the substation, the average value of the reactive power load in each period throughout the year is calculated, and the whole day period is divided into peak period, flat period and valley period according to the average value.

3. The method of claim 2, wherein the method is based on feature index dimension reduction of daily reactive load curve clustering. The daily reactive power load characteristic indexes are characteristic indexes in different periods, specifically: the load rate, the maximum utilization hour rate and the daily peak-valley difference rate are daily reactive power load characteristic indexes in the whole day period, the peak period load rate and the maximum load occurrence time are daily reactive power load characteristic indexes in the peak period, the flat period load rate is a daily reactive power load characteristic index in the flat period, and the valley period load rate and the minimum load occurrence time are daily reactive power load characteristic indexes in the valley period.

4. The method of claim 1, wherein the method is based on feature index dimension reduction for daily reactive load curve clustering. Expert weighting method is used to assign weights to daily reactive power load characteristic indexes, and the obtained daily reactive power load characteristic index weight matrix is: wherein W represents a daily reactive load characteristic index weight matrix, W m represents a daily reactive load characteristic index weight data set of the mth expert, d represents the number of invited experts, c represents the number of daily reactive load characteristic indexes, w ml represents the lth daily reactive load characteristic index weight of the mth expert, and satisfies 5. The method of claim 4, wherein the method is based on feature index dimension reduction of daily reactive load curve clustering. The characteristic index weight average of all experts is calculated based on the obtained daily reactive power load characteristic index weight matrix, and the characteristic index weight average is composed of the weight average assigned by all experts to each daily reactive power load characteristic index.

6. The method of claim 1, wherein the method is based on feature index dimension reduction for daily reactive load curve clustering. The error sum of squares SSE of the whole daily reactive power load data set is taken as the judgment condition for ending clustering.

7. The method of claim 1, wherein the method is based on feature index dimension reduction for daily reactive load curve clustering. After clustering is stopped when the judgment condition for ending clustering is met, the "elbow point" in the change curve of the error sum of squares SSE with the change of k value is taken as the best clustering number and the clustering result is outputted.

8. A clustering system for daily reactive load curves based on feature index dimensionality reduction, characterized in that, The method comprises the following steps: a data acquisition module configured to obtain annual reactive power load raw data of a substation; a characteristic index selection module configured to take load rate, maximum utilization hour rate, daily peak-valley difference rate, peak period load rate, maximum load occurrence time, flat period load rate, valley period load rate and minimum load occurrence time as daily reactive power load characteristic indexes; a data processing module configured to perform dimensionality reduction on daily reactive power load data based on daily reactive power load characteristic indexes to obtain a daily reactive power load characteristic index data reduction matrix, and to assign weights to daily reactive power load characteristic indexes using expert weighting method; a clustering module configured to perform clustering calculation on the dimensionally reduced daily reactive power load data based on k-means clustering algorithm, and to stop clustering and output a daily reactive power load curve clustering result when a judgment condition for ending clustering is met.

9. A computer-readable storage medium having stored thereon a program, characterized in that, The program is executed by the processor to realize the steps of the daily reactive power load curve clustering method based on characteristic index dimensionality reduction according to any one of claims 1-7.

10. An electronic device comprising a memory, a processor, and a program stored in the memory and capable of running on the processor, characterized by The processor implements the steps in the feature index dimension reduction based daily reactive load curve clustering method according to any one of claims 1-7 when executing the program.

Images