Power system user load data preprocessing method and system

Abstract

The invention discloses an electric power system user load data preprocessing method. The method comprises the steps of obtaining user original load data collected by an electric power system; calculating daily electric quantity and daily load data of the user; detecting load data; correcting the load data; clustering the corrected load data by combining a whale optimization algorithm and a clustering algorithm to obtain a load characteristic curve of the user; and finishing the final preprocessing of the user load data of the power system. The invention also discloses a system for realizing the user load data preprocessing method of the power system. According to the method, user load clustering is performed on the basis of processing the user load acquisition data in combination with the whale optimization algorithm and the clustering algorithm, so that preprocessing of the user load data of the power system is realized, the reliability is higher, and the accuracy is better.

Description

Technical Field

[0001] The invention belongs to the field of electrical automation, and in particular relates to a method and system for preprocessing user load data of an electric power system. Background Art

[0002] With the development of economy and technology and the improvement of people's living standards, electricity has become an indispensable secondary energy source in people's production and life, bringing endless convenience to people's production and life. Therefore, ensuring the stable and reliable supply of electricity has become one of the most important tasks of the power system.

[0003] Load management refers to improving the user's electricity consumption behavior and electricity usage mode through load adjustment technology, so as to reduce the maximum load of the power grid, save electricity, reduce the installed capacity of the power system, and reduce power investment and operating costs. With the rapid development of the power system, the load management of the power system is becoming more and more important.

[0004] In order to achieve refined load management, user load data of the power system is crucial. At present, user load data of the power system is generally based on power marketing data and power system collection data, and the power system collects load data through the route of collecting master station-substation terminal-smart energy meter. However, at present, the process of load data collection is easily interfered by abnormal factors such as weather, signals, and equipment failures; these abnormal factors will cause the collected load data to be distorted, which will seriously affect the load management of the power system. Summary of the invention

[0005] One of the purposes of the present invention is to provide a method for preprocessing user load data of a power system with high reliability and good accuracy.

[0006] A second objective of the present invention is to provide a system for implementing the power system user load data preprocessing method.

[0007] The power system user load data preprocessing method provided by the present invention comprises the following steps:

[0008] S1. Obtain the original load data of users collected by the power system;

[0009] S2. Calculate the user's daily power consumption and daily load data based on the original load data obtained in step S1;

[0010] S3. According to the load data obtained in step S2, the load data is detected;

[0011] S4. Correct the load data according to the test results obtained in step S3;

[0012] S5. Combine the whale optimization algorithm and the clustering algorithm to cluster the corrected load data to obtain the user's load characteristic curve;

[0013] S6. Complete the preprocessing of the final power system user load data.

[0014] The step S1 of obtaining the original user load data collected by the power system specifically includes the following steps:

[0015] Based on the data middle platform, the archive data of the power system marketing system is obtained; the archive data includes user archive data, metering point archive data and electric energy meter archive data;

[0016] Obtain daily frozen indication data and power curve data from the power system acquisition system.

[0017] The calculation of the user's daily power consumption and daily load data based on the original load data obtained in step S1 described in step S2 specifically includes the following steps:

[0018] The following formula is used to calculate the daily electricity consumption of the metering point:

[0019] Q=(SS t -SS t-1 ) · PT · CT

[0020] Where Q is the daily electricity consumption at the metering point; SS t is the frozen value on day t; SS t-1 is the frozen indication on day t-1; PT is the current transformation ratio; CT is the voltage transformation ratio;

[0021] The following formula is used to calculate the momentary load data of the metering point:

[0022] L=p·PT·CT

[0023] Where L is the load data at the metering point at the moment; p is the power data at the moment;

[0024] The user's daily electricity data is the sum of the electricity at the user's metering points;

[0025] The user's daily load data is the sum of the loads at the user's metering points.

[0026] The step S3 of detecting the load data according to the load data obtained in step S2 specifically includes the following steps:

[0027] The load data obtained in step S2 is traversed, and anomaly detection is performed using the following rules:

[0028] Rule 1: If the load data at a certain point is greater than X times the user's contract capacity or greater than Y times the maximum power of the electric energy meter, the load data at that point is judged to be abnormally large;

[0029] Rule 2: If the load data of a certain point is empty, the load data of this point is judged to be missed;

[0030] Rule 3: If the load data of a certain point is negative, the load data of the point is judged to be negative value abnormal;

[0031] Record the number of consecutive abnormal moments of excessive number anomalies and missed sampling anomalies.

[0032] Step S4, in accordance with the detection result obtained in step S3, corrects the load data, and specifically includes the following steps:

[0033] For load data with abnormal negative values:

[0034] Directly set the corresponding load data to 0 to correct the abnormal negative load data;

[0035] For load data with abnormally large numbers and missed data:

[0036] If the number of consecutive abnormal moments is less than or equal to the set threshold T1, the following formula is used for correction:

[0037] Continuous abnormal data is represented by [P n ',P n ' +1 ,...,P n ' +k ], k is the number of consecutive abnormal data, and k≤T1-1; the following formula is used for correction:

[0038]

[0039] Where P n+i is the i+1th abnormal data after repair; P n ' +i+1 is the n+i+2th abnormal data before repair; P n ' -1 is the previous data of the continuous abnormal data;

[0040] If the number of consecutive abnormal moments is greater than the set threshold T1 and less than the set threshold T2, T2>T1, then for the consecutive abnormal data, the average value of the load data of the user corresponding to the data in the previous few days is calculated, and the obtained load data average value is used as the correction value of the consecutive abnormal data;

[0041] If the number of consecutive abnormal moments is greater than or equal to the set threshold T2, the power and baseline load data of the user corresponding to the data are obtained, and the user's daily power Q and user baseline load {J1, J2, ..., Jn} are calculated, where n is the number of load points; the user's normal load data is represented by Pi , abnormal data is represented by P i ' ？ ; Calculate the user's normal load fitting power Q zc Q zc =∑P i ×0.25, and then calculate the correction value P of the abnormal data i？ for Where Ji' is the user baseline load value corresponding to the abnormal data moment.

[0042] Step S5 combines the whale optimization algorithm and the clustering algorithm to cluster the corrected load data to obtain the user's load characteristic curve, which specifically includes the following steps:

[0043] A. Initialize the number of clusters K of the K-means clustering algorithm; randomly initialize K cluster centers, and set the population size and maximum number of iterations of the whale optimization algorithm;

[0044] B. The whale optimization algorithm randomly generates a set of potential centroid locations as the starting positions of the search agent; the whale searches for better cluster center locations by surrounding prey, attacking with bubble nets, and searching for prey;

[0045] C. Use the updated whale position obtained by the whale optimization algorithm as the initial cluster center of K-means; execute the K-means clustering algorithm, redistribute the data points to the nearest cluster center, and update the position of the cluster center;

[0046] D. Repeat step C until the cluster center no longer changes or the set number of repetitions is reached;

[0047] E. Based on the final cluster center and the corresponding cluster curve results, the user's load characteristic curve is obtained.

[0048] The encirclement of the prey in step B specifically comprises the following steps:

[0049] Assume that in d-dimensional space, the current best whale individual X * The location is Whale Individual X j The location is

[0050] Whale Individual X j Next position under the influence of the best individual whales Expressed as

[0051]

[0052] In the formula For X j+1 The k-th dimension element value of the position; is the current best whale individual X * The k-th dimension element value of the position; A 1 is the first intermediate variable, and A 1 =2α·γ 1 -α, α is the number of changes from 2 to 0 as the number of iterations increases, γ 1 is the first random number ranging from 0 to 1; D k is the second intermediate variable, and C 1 is the third intermediate variable and C 1 =2γ 2 , γ 2 is a second random number ranging from 0 to 1.

[0053] The bubble network attack described in step B specifically includes the following steps:

[0054] Assume that in d-dimensional space, the current best whale individual X * The location is Whale Individual X j The location is

[0055] Shrink wrap position update:

[0056] Whale Individual X j Next position under the influence of the best individual whales Expressed as

[0057]

[0058] Where A 1 ' is the fourth intermediate variable, and its value range is [-1,1];

[0059] Spiral Position Update:

[0060] Whale Individual X j Next position under the influence of the best individual whales Expressed as

[0061]

[0062] Where E k is the fifth intermediate variable, and b is a logarithmic spiral shape constant; l is a third random number ranging from -1 to 1;

[0063] During the bubble net attack, individual whales choose to update their positions in a shrinking and encircling manner or in a spiral manner with a set probability.

[0064] The search and predation described in step B specifically includes the following steps:

[0065] Suppose in d-dimensional space, a random whale individual X in the current whale group rand The location is Whale Individual X j The location is

[0066] The following formula is used to represent the position update of the search and prey process:

[0067]

[0068] In the formula is a random whale individual X rand The k-th dimension element value at the position; D' k is the sixth intermediate variable, and

[0069] The present invention also provides a system for implementing the power system user load data preprocessing method, comprising a data acquisition module, a data calculation module, a data detection module, a data correction module, a data clustering module and a data preprocessing module; the data acquisition module, the data calculation module, the data detection module, the data correction module, the data clustering module and the data preprocessing module are connected in series in sequence; the data acquisition module is used to acquire the user's original load data collected by the power system, and upload the data information to the data calculation module; the data calculation module is used to calculate the user's daily power consumption and daily load data based on the received data information and the acquired original load data, and upload the data information to the data detection module. Module; the data detection module is used to detect the load data according to the received data information and the obtained load data, and upload the data information to the data correction module; the data correction module is used to correct the load data according to the received data information and the obtained detection results, and upload the data information to the data clustering module; the data clustering module is used to cluster the corrected load data according to the received data information, combined with the whale optimization algorithm and the clustering algorithm, to obtain the user's load characteristic curve, and upload the data information to the data preprocessing module; the data preprocessing module is used to complete the final preprocessing of the power system user load data according to the received data information.

[0070] The power system user load data preprocessing method and system provided by the present invention are based on the processing of user load collection data and the user load clustering in combination with the whale optimization algorithm and the clustering algorithm, which not only realizes the preprocessing of power system user load data, but also has higher reliability and better accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 The figure is a schematic diagram of the method flow of the present invention.

[0072] Figure 2 Schematic diagram of the functional modules of the system of the present invention. DETAILED DESCRIPTION

[0073] like Figure 1 The method flow chart of the method of the present invention is shown as follows: The power system user load data preprocessing method disclosed in the present invention comprises the following steps:

[0074] S1. Obtaining the original user load data collected by the power system; specifically including the following steps:

[0075] Based on the data middle platform, the archive data of the power system marketing system is obtained; the archive data includes user archive data, metering point archive data and electric energy meter archive data;

[0076] Obtain daily frozen indication data and power curve data from the power system acquisition system;

[0077] S2. Based on the original load data obtained in step S1, the user's daily electricity consumption and daily load data are calculated; specifically, the following steps are included:

[0078] The following formula is used to calculate the daily electricity consumption of the metering point:

[0079] Q=(SS t -SS t-1 ) · PT · CT

[0080] Where Q is the daily electricity consumption at the metering point; SS t is the frozen value on day t; SS t-1 is the frozen indication on day t-1; PT is the current transformation ratio; CT is the voltage transformation ratio;

[0081] The following formula is used to calculate the momentary load data of the metering point:

[0082] L=p·PT·CT

[0083] Where L is the load data at the metering point at the moment; p is the power data at the moment;

[0084] The user's daily electricity data is the sum of the electricity at the user's metering points;

[0085] The daily load data of a user is the sum of the loads at the user's metering points;

[0086] S3. According to the load data obtained in step S2, the load data is detected; specifically comprising the following steps:

[0087] The load data obtained in step S2 is traversed, and anomaly detection is performed using the following rules:

[0088] Rule 1: If the load data at a certain point is greater than X times (preferably 5 times) the user's contract capacity or greater than Y times (preferably 2 times) the maximum power of the electric energy meter, the load data at this point is judged to be abnormally large;

[0089] Rule 2: If the load data of a certain point is empty, the load data of this point is judged to be missed;

[0090] Rule 3: If the load data of a certain point is negative, the load data of the point is judged to be negative value abnormal;

[0091] Record the number of consecutive abnormal moments of excessive number anomalies and missed sampling anomalies;

[0092] S4. According to the detection result obtained in step S3, the load data is corrected; specifically comprising the following steps:

[0093] For load data with abnormal negative values:

[0094] Directly set the corresponding load data to 0 to correct the abnormal negative load data;

[0095] For load data with abnormally large numbers and missed data:

[0096] If the number of consecutive abnormal moments is less than or equal to the set threshold T1 (preferably 6), the following formula is used for correction:

[0097] Continuous abnormal data is represented by [P n ',P n ' +1 ,...,P n ' +k ], k is the number of consecutive abnormal data, and k≤T1-1; the following formula is used for correction:

[0098]

[0099] Where P n+i is the i+1th abnormal data after repair; P n ' +i+1 is the n+i+2th abnormal data before repair; P n ' -1 is the previous data of the continuous abnormal data;

[0100] If the number of consecutive abnormal moments is greater than the set threshold T1 and less than the set threshold T2 (preferably 48), T2>T1, then for the consecutive abnormal data, the average value of the load data of the user corresponding to the data in the previous few days is calculated, and the obtained load data average value is used as the correction value of the consecutive abnormal data;

[0101] If the number of consecutive abnormal moments is greater than or equal to the set threshold T2, the power and baseline load data of the user corresponding to the data are obtained, and the user's daily power Q and user baseline load {J1, J2, ..., Jn} are calculated, where n is the number of load points; the user's normal load data is represented by P i , abnormal data is represented by P i ' ？ ; Calculate the user's normal load fitting power Q zc Q zc =∑P i ×0.25, and then calculate the correction value P of the abnormal data i？ for Where Ji' is the user baseline load value corresponding to the abnormal data moment;

[0102] S5. Combining the whale optimization algorithm and the clustering algorithm, clustering the corrected load data to obtain the user's load characteristic curve; specifically, the steps include:

[0103] A. Initialize the number of clusters K of the K-means clustering algorithm; randomly initialize K cluster centers, and set the population size and maximum number of iterations of the whale optimization algorithm;

[0104] In the specific implementation, the K-means clustering algorithm is used to generate a data sample library from the load data objects of a single user over a period of time, and the sample data is classified according to the season (spring, summer, autumn and winter) and date type (working day, holiday); the clustering algorithm must check whether the classification of each sample is correct in each iteration, if not, it must be adjusted; after all samples are adjusted, the cluster center is modified and the next iteration is entered; if all samples are correctly classified in one iteration of the algorithm, there will be no adjustment and the cluster center will not change; the value is constantly decreasing in the algorithm iteration and finally converges to a fixed value;

[0105] B. The whale optimization algorithm randomly generates a set of potential centroid locations as the starting positions of the search agent; the whale searches for better cluster center locations by surrounding prey, attacking with bubble nets, and searching for prey;

[0106] Since the traditional K-means clustering algorithm has certain limitations, such as being greatly affected by the centroid of the initial cluster, and since the K-means algorithm is an iterative process based on gradient descent, it is easy to fall into local optimality rather than global optimality; therefore, in order to improve the shortcomings of the K-means clustering algorithm, the scheme of the present invention combines the whale optimization algorithm with K-means; using the global search capability of the whale optimization algorithm to improve the selection of the initial clustering center of K-means, it helps to avoid local optimal solutions, thereby achieving the purpose of improving the clustering effect and stability of the algorithm;

[0107] The whale optimization algorithm is a new swarm intelligence optimization algorithm that imitates the predation behavior of whales in nature. The predation behavior of whales can be divided into three categories: 1) surrounding prey; 2) attacking with a bubble net; 3) searching for prey;

[0108] The purpose of whale hunting is to capture prey. When a group of whales are looking for prey together, there will definitely be a situation where one whale finds the prey first. At this time, other whales will swim towards the whale that has found the prey to compete for the prey. This process is applied specifically, that is, one solution can be represented by one whale individual, and several solutions can be represented by several whale individuals. The process of searching for a solution to a problem using the whale optimization algorithm can be regarded as several whale individuals constantly updating their individual positions until a satisfactory solution is found.

[0109] In specific implementation, the encirclement of the prey comprises the following steps:

[0110] Assume that in d-dimensional space, the current best whale individual X * The location is Whale Individual X j The location is

[0111] Whale Individual X j Next position under the influence of the best individual whales Expressed as

[0112]

[0113] In the formula For X j+1 The k-th dimension element value of the position; is the current best whale individual X * The k-th dimension element value of the position; A 1 is the first intermediate variable, and A 1 =2α·γ 1 -α, α is the number of changes from 2 to 0 as the number of iterations increases, γ 1 is the first random number ranging from 0 to 1; D k is the second intermediate variable, and C 1 is the third intermediate variable and C 1 =2γ 2 , γ 2 is a second random number ranging from 0 to 1;

[0114] The bubble network attack specifically includes the following steps:

[0115] Assume that in d-dimensional space, the current best whale individual X * The location is Whale Individual X j The location is

[0116] Shrink wrap position update:

[0117] Whale Individual X j Next position under the influence of the best individual whales Expressed as

[0118]

[0119] Where A 1 ' is the fourth intermediate variable, and its value range is [-1,1];

[0120] Spiral Position Update:

[0121] Whale Individual X j Next position under the influence of the best individual whales Expressed as

[0122]

[0123] Where E k is the fifth intermediate variable, and b is a logarithmic spiral shape constant; l is a third random number ranging from -1 to 1;

[0124] During the bubble net attack, individual whales choose to update their positions in a shrinking and encircling manner or in a spiral manner with a set probability;

[0125] The search and predation method specifically comprises the following steps:

[0126] In the mathematical model of the shrinking predator behavior, A 1 ' is limited to [-1,1]; but when A 1 When the value of ' is not in [-1,1], the current individual whale may not approach the current best individual whale, but randomly select a whale from the current whale group to approach. This is the idea of ​​search and predation. Although search and predation may cause the current individual whale to deviate from the target prey, it will enhance the global search ability of the whale group.

[0127] Suppose in d-dimensional space, a random whale individual X in the current whale group rand The location is Whale Individual X j The location is

[0128] The following formula is used to represent the position update of the search and prey process:

[0129]

[0130] In the formula is a random whale individual X rand The k-th dimension element value at the position; D' k is the sixth intermediate variable, and

[0131] C. Use the updated whale position obtained by the whale optimization algorithm as the initial cluster center of K-means; execute the K-means clustering algorithm, redistribute the data points to the nearest cluster center, and update the position of the cluster center;

[0132] D. Repeat step C until the cluster center no longer changes or the set number of repetitions is reached;

[0133] E. Obtain the user's load characteristic curve based on the final cluster center and the corresponding cluster curve result;

[0134] S6. Complete the preprocessing of the final power system user load data; finally, after completing the data preprocessing according to the above steps, it is possible to generate eight types of clustered load curves for power users: spring working days, spring non-working days, summer working days, summer non-working days, autumn working days, autumn non-working days, winter working days, and winter non-working days.

[0135] The generated load curve can be used for subsequent work, such as user load feature identification: load features include maximum load, minimum load, peak-valley difference, peak-valley difference rate, daily load rate, minimum load rate, etc. Among them, peak-valley difference = peak section maximum load - valley section minimum load, peak-valley difference rate = peak-valley difference / daily maximum load, daily load rate = daily average load / daily maximum load, and minimum load rate = daily minimum load / daily maximum load.

[0136] like Figure 2The functional module diagram of the system of the present invention is shown as follows: the system disclosed in the present invention for implementing the method for preprocessing the user load data of the power system comprises a data acquisition module, a data calculation module, a data detection module, a data correction module, a data clustering module and a data preprocessing module; the data acquisition module, the data calculation module, the data detection module, the data correction module, the data clustering module and the data preprocessing module are connected in series in sequence; the data acquisition module is used to acquire the original load data of the user collected by the power system, and upload the data information to the data calculation module; the data calculation module is used to calculate the daily power consumption and daily load data of the user based on the received data information and the acquired original load data, and upload the data to the data calculation module; Information upload data detection module; the data detection module is used to detect the load data according to the received data information and the obtained load data, and upload the data information to the data correction module; the data correction module is used to correct the load data according to the received data information and the obtained detection results, and upload the data information to the data clustering module; the data clustering module is used to cluster the corrected load data according to the received data information, combined with the whale optimization algorithm and the clustering algorithm, to obtain the user's load characteristic curve, and upload the data information to the data preprocessing module; the data preprocessing module is used to complete the final preprocessing of the power system user load data according to the received data information.

Claims

1. A method for preprocessing load data of power system users, comprising the following steps: S1. Obtain the original load data of users collected by the power system; S2. Calculate the user's daily power consumption and daily load data based on the original load data obtained in step S1; S3. According to the load data obtained in step S2, the load data is detected; S4. Correct the load data according to the test results obtained in step S3; S5. Combine the whale optimization algorithm and the clustering algorithm to cluster the corrected load data to obtain the user's load characteristic curve; S6. Complete the preprocessing of the final power system user load data.

2. The method for preprocessing user load data of a power system according to claim 1, characterized in that The step S1 of obtaining the original user load data collected by the power system specifically includes the following steps: Based on the data middle platform, the archive data of the power system marketing system is obtained; the archive data includes user archive data, metering point archive data and electric energy meter archive data; Obtain daily frozen indication data and power curve data from the power system acquisition system.

3. The method for preprocessing user load data of a power system according to claim 2, characterized in that The calculation of the user's daily power consumption and daily load data based on the original load data obtained in step S1 described in step S2 specifically includes the following steps: The following formula is used to calculate the daily electricity consumption of the metering point: Q=(SS t -SS t-1 )·PT·CT Where Q is the daily electricity consumption at the metering point; SS t is the frozen value on day t; SS t-1 is the frozen indication on day t-1; PT is the current transformation ratio; CT is the voltage transformation ratio; The following formula is used to calculate the momentary load data of the metering point: L=p·PT·CT Where L is the load data at the metering point at the moment; p is the power data at the moment; The user's daily electricity data is the sum of the electricity at the user's metering points; The user's daily load data is the sum of the loads at the user's metering points.

4. The method for preprocessing user load data of a power system according to claim 3, characterized in that The step S3 of detecting the load data according to the load data obtained in step S2 specifically includes the following steps: The load data obtained in step S2 is traversed, and anomaly detection is performed using the following rules: Rule 1: If the load data at a certain point is greater than X times the user's contract capacity or greater than Y times the maximum power of the electric energy meter, the load data at that point is considered to be abnormally large; Rule 2: If the load data of a certain point is empty, the load data of this point is judged to be missed; Rule 3: If the load data of a certain point is negative, the load data of the point is judged to be negative value abnormal; Record the number of consecutive abnormal moments of excessive number anomalies and missed sampling anomalies.

5. The method for preprocessing user load data of a power system according to claim 4, characterized in that Step S4, in accordance with the detection result obtained in step S3, corrects the load data, and specifically includes the following steps: For load data with abnormal negative values: Directly set the corresponding load data to 0 to correct the abnormal negative load data; For load data with abnormally large numbers and missed data: If the number of consecutive abnormal moments is less than or equal to the set threshold T1, the following formula is used for correction: Continuous abnormal data is represented by [P n ',P n ' +1 ,...,P n ' +k ], k is the number of consecutive abnormal data, and k≤T1-1; the following formula is used for correction: Where P n+i is the i+1th abnormal data after repair; P n ' +i+1 is the n+i+2th abnormal data before repair; P n ' -1 is the previous data of the continuous abnormal data; If the number of consecutive abnormal moments is greater than the set threshold T1 and less than the set threshold T2, T2>T1, then for the consecutive abnormal data, the average value of the load data of the user corresponding to the data in the previous few days is calculated, and the obtained load data average value is used as the correction value of the consecutive abnormal data; If the number of consecutive abnormal moments is greater than or equal to the set threshold T2, the power and baseline load data of the user corresponding to the data are obtained, and the user's daily power Q and user baseline load {J1, J2, ..., Jn} are calculated, where n is the number of load points; the user's normal load data is represented by P i , abnormal data is represented by P i ' ？ ; Calculate the user's normal load fitting power Q zc Q zc =∑P i ×0.25, and then calculate the correction value P of the abnormal data i？ for Where Ji' is the user baseline load value corresponding to the abnormal data moment.

6. The method for preprocessing user load data of a power system according to claim 5, characterized in that Step S5 combines the whale optimization algorithm and the clustering algorithm to cluster the corrected load data to obtain the user's load characteristic curve, which specifically includes the following steps: A. Initialize the number of clusters K of the K-means clustering algorithm; randomly initialize K cluster centers, and set the population size and maximum number of iterations of the whale optimization algorithm; B. The whale optimization algorithm randomly generates a set of potential centroid locations as the starting positions of the search agent; the whale searches for better cluster center locations by surrounding prey, attacking with bubble nets, and searching for prey; C. Use the updated whale position obtained by the whale optimization algorithm as the initial cluster center of K-means; execute the K-means clustering algorithm, redistribute the data points to the nearest cluster center, and update the position of the cluster center; D. Repeat step C until the cluster center no longer changes or the set number of repetitions is reached; E. Based on the final cluster center and the corresponding cluster curve results, the user's load characteristic curve is obtained.

7. The method for preprocessing user load data of a power system according to claim 6, characterized in that The encirclement of the prey described in step B specifically comprises the following steps: Assume that in d-dimensional space, the current best whale individual X * The location is Whale Individual X j The location is Whale Individual X j Next position under the influence of the best individual whales Expressed as In the formula For X j+1 The k-th dimension element value of the position; is the current best whale individual X * The k-th dimension element value at the position of ; A1 is the first intermediate variable, and A1 = 2α·γ1-α, α is the change number that decreases from 2 to 0 as the number of iterations increases, and γ1 is the first random number with a value range of 0 to 1; D k is the second intermediate variable, and C1 is the third intermediate variable and C1=2γ2, where γ2 is a second random number ranging from 0 to 1.

8. The method for preprocessing user load data of a power system according to claim 7, characterized in that The bubble network attack described in step B specifically includes the following steps: Assume that in d-dimensional space, the current best whale individual X * The location is Whale Individual X j The location is Shrink wrap position update: Whale Individual X j Next position under the influence of the best individual whales Expressed as Where A1' is the fourth intermediate variable, and its value range is [-1,1]; Spiral Position Update: Whale Individual X j Next position under the influence of the best individual whales Expressed as Where E k is the fifth intermediate variable, and b is a logarithmic spiral shape constant; l is a third random number ranging from -1 to 1; During the bubble net attack, individual whales choose to update their position in a shrinking and encircling manner or in a spiral manner with a set probability.

9. The method for preprocessing user load data of a power system according to claim 8, characterized in that The search and predation described in step B specifically includes the following steps: Suppose in d-dimensional space, a random whale individual X in the current whale group rand The location is Whale Individual X j The location is The following formula is used to represent the position update of the search and prey process: In the formula For a random whale individual X rand The k-th dimension element value at the position; D' k is the sixth intermediate variable, and 10. A system for implementing the method for preprocessing power system user load data according to any one of claims 1 to 9, characterized in that It includes a data acquisition module, a data calculation module, a data detection module, a data correction module, a data clustering module and a data preprocessing module; the data acquisition module, the data calculation module, the data detection module, the data correction module, the data clustering module and the data preprocessing module are connected in series in sequence; the data acquisition module is used to obtain the original load data of the user collected by the power system, and upload the data information to the data calculation module; The data calculation module is used to calculate the user's daily power consumption and daily load data according to the received data information and the acquired original load data, and upload the data information to the data detection module; The data detection module is used to detect the load data according to the received data information and the obtained load data, and upload the data information to the data correction module; The data correction module is used to correct the load data according to the received data information and the obtained detection results, and upload the data information to the data clustering module; The data clustering module is used to cluster the corrected load data according to the received data information, combined with the whale optimization algorithm and the clustering algorithm, to obtain the user's load characteristic curve, and upload the data information to the data preprocessing module; The data preprocessing module is used to complete the preprocessing of the final power system user load data based on the received data information.

Images