A method, apparatus, equipment, and medium for high-voltage user curve fitting based on a typical industry feature library.

Abstract

This invention discloses a method, apparatus, equipment, and medium for high-voltage user curve fitting based on a typical industry feature library. The method includes: constructing an industry feature library based on high-voltage user profile data and electricity consumption curve data; obtaining a reference curve; fitting the current electricity consumption curve with the obtained reference curve, replacing missing electricity consumption values; verifying the replaced values, and determining whether the replacement method needs to be changed based on the verification results. This high-voltage user curve fitting method, apparatus, equipment, and medium based on an industry feature library utilizes customer profile data and electricity consumption curve data accumulated by a new generation of electricity consumption data acquisition and power marketing systems. Through big data analysis, it effectively completes missing values ​​in the electricity consumption curve, accurately estimates user electricity consumption, and protects the interests of power supply companies and large customers.

Description

Technical Field

[0001] This invention relates to a method, apparatus, equipment, and medium for high-voltage user curve fitting based on a typical industry feature library, belonging to the field of electricity data analysis and processing technology. Background Technology

[0002] With the continuous development and improvement of the new generation of electricity information collection systems, it is now possible to collect 15-minute-level power consumption curve data for high-voltage users. These curves are an important basis for monitoring users' electricity consumption and calculating electricity usage. However, due to unstable network signals and equipment malfunctions, the power consumption values ​​at some points in time cannot be properly transmitted back to the collection system, resulting in incomplete power consumption curves. Consequently, it is impossible to accurately settle the electricity bills for high-voltage users at different times and predict their electricity demand.

[0003] It is evident that, in order to solve the problem of incomplete power consumption curves and to complete them, there is an urgent need for a high-voltage user curve fitting method, device, equipment, and medium based on an industry feature library. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method, device, equipment and medium for high-voltage user curve fitting based on a typical industry feature library. By utilizing customer profile data and power consumption curve data accumulated by the new generation of electricity consumption information system and power marketing system, and through big data analysis, the missing values ​​in the power consumption curve are effectively supplemented, thereby accurately estimating the user's electricity consumption and protecting the interests of the power supply company and the user.

[0005] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0006] In a first aspect, the present invention provides a method for high-voltage user curve fitting based on a typical industry feature library, comprising:

[0007] An industry feature library was constructed based on the archive data of high-voltage users and the power consumption indicator curve data.

[0008] Acquire user data with missing energy readings, categorize the daily energy reading curve data with missing points, calculate the similarity with the cluster center curves in the industry feature library using standardized Euclidean distance, and obtain reference curves;

[0009] The current daily energy consumption reading curve is fitted with the acquired reference curve, and the missing energy consumption readings are replaced.

[0010] Perform data validation on the replacement values, and determine whether the replacement method needs to be changed based on the validation results.

[0011] Furthermore, the method for constructing the industry feature library includes:

[0012] Acquire the profile data and power consumption curve data of high-voltage users;

[0013] Based on the acquired archival data, the dimensions that affect the electricity data in the existing archival data are extracted and classified. The classified dimensions include industry, region and electricity consumption level.

[0014] Obtain the daily energy consumption reading curve data for each category, and segment and classify the curve data according to the date corresponding to the curve data. The classification dimensions include season, weather and temperature.

[0015] Cluster the daily electricity consumption curves;

[0016] The clusters obtained from clustering are named according to the feature dimensions, and an industry feature library is constructed based on industry, region, electricity consumption level, season, weather, temperature, and daily energy consumption reading curve.

[0017] Furthermore, the clustering method includes:

[0018] Data processing is performed on the daily electricity consumption indicator curve to calculate the electricity consumption increase data between two adjacent minimum intervals;

[0019] The price increase data was standardized using the Z-Score standardization method.

[0020] Based on the standardized increase data, the Euclidean distance formula is used to calculate the curve similarity.

[0021] An adaptive K-means algorithm is used to cluster the curves.

[0022] Furthermore, the method for clustering curves using the adaptive K-means algorithm includes:

[0023] Step 1: Determine the maximum number of clusters k max and minimum cluster number k min ;

[0024] Step 2: Let k = k min ;

[0025] Step 3: Cluster the power consumption curves of high-voltage users by calculating the distance between the curves using Euclidean distance based on the cluster number k;

[0026] Step 4: Calculate the evaluation index using the profile coefficient. The value range is [-1, 1];

[0027] Step 5: Increment k by 1, if k is not greater than k maxReturn to step three and continue clustering, performing steps three and four until k = k max No further clustering, proceed to step six;

[0028] Step 6: Compare the I values ​​of each clustering result. DBI , will the maximum value (I) DBI ) max The corresponding number of cluster centers k is used as the result of clustering.

[0029] Furthermore, the method for obtaining the reference curve includes:

[0030] Match industry feature databases based on user profile data;

[0031] Based on the missing points, process the daily energy consumption curve and calculate the electricity consumption increase data between two adjacent minimum intervals;

[0032] The Z-Score was used to standardize the price increase data;

[0033] Based on standardized growth rate data, the Euclidean distance between the current curve and each cluster center is calculated to obtain a reference curve;

[0034] Specifically as follows:

[0035] Calculate the mean and standard deviation of the curve X to be fitted, and then calculate the standardized value of each data point using the Z-Score standardization formula. The formula is as follows:

[0036]

[0037] In the formula, r xi ′ represents the standardized increase data, r xi This is the original increase data, μ is the average increase of the curve, and σ is the standard deviation of the increase data of the curve;

[0038] Based on standardized increase data, the curve to be fitted, X, and the cluster center, Y{y1,y2,…,y} are calculated. k The Euclidean distance to} is calculated using the following formula:

[0039]

[0040] In the formula, r xi ′ represents the i-th standardized increase data on the curve X to be fitted, r yi ′ refers to the i-th standardized increase data on curve Y, and d(X,Y) represents the similarity between curve X and curve Y;

[0041] Compare k d(X,Y) values ​​and find the minimum Euclidean distance (d(X,Y)). minThe corresponding cluster center Y′ is used as the reference curve for the curve X to be fitted.

[0042] Furthermore, the method for replacing the electrical energy readings at the missing points includes:

[0043] Obtain the energy readings at 97 points on the reference curve Y′, and calculate the energy consumption increase {r1, r2, ... r} between 95 adjacent minimum intervals. 95}

[0044] Obtain the missing points on the curve X to be fitted, mark the time series of the missing points, sort the missing points according to the time series, calculate the energy readings of the missing points in turn, and use the calculated energy readings as replacement values. The formula for calculating the energy readings of the missing points is as follows:

[0045]

[0046] In the above formula, q i The energy reading is the time series of missing points on curve X, where q is the energy reading. i-1 q i+1 It is a missing point q i The electrical energy readings at the two points before and after, r i-1 It represents the increase in the reference curve and the time series corresponding to the missing points.

[0047] Furthermore, the data verification method includes:

[0048] The Grubbs criterion was used to identify outliers, and the electricity readings {q1,q2,…,q} of 97 points on the day the missing point occurred were obtained. 97 Arrange the data in descending order and count q1 <... 97 The standard deviation is calculated using the following formula:

[0049]

[0050] The Grubbs value is calculated using the following definition:

[0051]

[0052] In the formula, μ is q i The average value is g i For Grubbs values, σ is the average of the 97 measurements, and σ is the standard deviation of the 97 values.

[0053] With a confidence probability P = 0.95 and a critical value g... p (97) = 3.196, if g i ​If the value is ≥3.196, then i is an outlier and needs to be removed and replaced by other methods. These other methods include replacing the original fitted result with the average value of the missing value over the previous 3 days.

[0054] Secondly, the present invention provides a high-voltage user curve fitting device based on a typical industry feature library, comprising:

[0055] Industry Feature Library Construction Module: Used to construct an industry feature library based on high-voltage user profile data and power consumption indicator curve data;

[0056] Reference curve acquisition module: used to acquire user data with missing power readings, classify the daily power reading curve data with missing points, calculate the similarity with the cluster center curves in the industry feature library using standardized Euclidean distance, and obtain reference curves;

[0057] Energy reading replacement module: This module is used to fit the current daily energy reading curve with the acquired reference curve and replace the missing energy readings.

[0058] Data validation module: Used to validate the replacement value and determine whether the replacement method needs to be changed based on the validation result.

[0059] Thirdly, the present invention provides a computer device, including a processor and a storage medium;

[0060] The storage medium is used to store instructions;

[0061] The processor is configured to operate according to the instructions to perform the steps of the method according to the first aspect.

[0062] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that the program, when executed by a processor, implements the steps of the method described in the first aspect.

[0063] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0064] The present invention provides a high-voltage user curve fitting method, device, equipment and medium based on a typical industry feature library. It utilizes customer profile data and power consumption curve data accumulated by the new generation of electricity consumption information system and power marketing system, and effectively fills in the missing values ​​in the power consumption curve through big data analysis, thereby accurately estimating the user's electricity consumption and protecting the interests of power supply companies and users. Attached Figure Description

[0065] Figure 1 This is a schematic diagram illustrating the steps of high-voltage user curve fitting based on an industry feature library.

[0066] Figure 2 This is a schematic diagram of the process for high-voltage user curve fitting based on an industry feature library. Detailed Implementation

[0067] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments and specific features in the embodiments are detailed descriptions of the technical solution of the present application, rather than limitations thereof. In the absence of conflict, the embodiments and technical features in the embodiments can be combined with each other.

[0068] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0069] Example 1:

[0070] Figure 1 This is a flowchart of a high-voltage user curve fitting method based on a typical industry feature library, as described in Embodiment 1 of the present invention. The high-voltage user curve fitting method based on a typical industry feature library provided in this embodiment can be applied to a terminal and can be executed by a high-voltage user curve fitting device based on the typical industry feature library. This device can be implemented in software and / or hardware and can be integrated into the terminal, such as any smartphone, tablet computer, or computer device with communication capabilities. See also... Figure 1 The method implemented in this way specifically includes the following steps:

[0071] Step A: Construct an industry feature library based on high-voltage user archive data, meteorological data, and power consumption indicator curve data;

[0072] Step B: Obtain user data with missing electricity readings, categorize the daily electricity reading curve data with missing points, match them with the industry feature library, calculate the similarity with the cluster center curves in the industry feature library using standardized Euclidean distance, and obtain the reference curve;

[0073] Step C: Fit the current daily energy reading curve with the acquired reference curve and replace the missing energy readings;

[0074] Step D: Perform data validation on the replacement value, and determine whether the replacement method needs to be changed based on the validation results.

[0075] Step Aa: The method for constructing the industry feature library includes:

[0076] First, based on the marketing business application system and the electricity information collection system, obtain the high-voltage user profile data and electricity reading curve data. The data source must be real and the curve must be complete.

[0077] Based on the acquired archival data, dimensions that influence electricity data were extracted from the existing archival data and categorized. The categorized dimensions include industry, region, and electricity consumption level, as detailed in the table below:

[0078] Dimension Dimensional classification Industry Classification Agriculture, forestry, animal husbandry and fishery, industry, finance, construction, real estate and other industries area Province + City + District / County voltage level 220V~380V, 1KV~10KV, 35KV~110KV...

[0079] Obtain the daily energy consumption indicator curve data for each category, and segment and categorize the curve data according to the corresponding date. The categorization dimensions include season, weather, and temperature, as shown in the table below:

[0080]

[0081] Cluster the daily electricity consumption curves;

[0082] The clusters obtained from clustering are named according to the feature dimensions, and an industry feature library is constructed based on industry, region, electricity consumption level, season, weather, temperature, and daily energy consumption curve. The establishment of a typical industry electricity consumption feature library is an important method to explore the differences, regularities, correlations, and trends of electricity theft characteristics in typical industries. By constructing an industry feature library, the electricity consumption patterns of similar users can be identified.

[0083] Step Ab: The clustering method includes:

[0084] Data processing was performed on the daily energy consumption reading curve. Values ​​were recorded at 15-minute intervals from 00:00 to 24:00, resulting in 97 points {q1,q2,…,q} per day. 97}, 96 minimum intervals {o1,o2,…,o 93} Calculate the electricity consumption increase data R{r1,r2,…,r between two adjacent minimum intervals. 95 Its definition is as follows:

[0085]

[0086] Z-Score standardization is used to standardize the daily electricity consumption readings. Applying Z-Score standardization to these readings eliminates differences in amplitude between different time points, allowing for a more accurate comparison of their similarity. Specifically, for each day's electricity consumption readings, the mean and standard deviation for that day are first calculated. Then, the standardized value for each data point is calculated using the Z-Score standardization formula. The specific formula is as follows:

[0087]

[0088] In the above formula, r i ′ refers to the standardized growth rate data, r i σ refers to the i-th increase data on the curve, μ refers to the average of all increase data on the curve, and σ refers to the variance of all increase data on the curve.

[0089] Curve similarity is calculated using the Euclidean distance formula. Euclidean distance is frequently used to calculate the distance between data points for data classification in clustering algorithms. The method is based on standardized price increase data, calculating the similarity between every two curves. The specific formula is as follows:

[0090]

[0091] In the above formula, r′ ai This refers to the i-th standardized increase data on curve A, r′ bi This refers to the i-th standardized increase data on curve B, where d(A,B) represents the similarity between curve A and curve B.

[0092] An adaptive K-means algorithm is used to cluster curves. Adaptive K-Means is an improved K-Means clustering algorithm that adaptively adjusts the number of clusters based on the silhouette coefficient evaluation metric, thus obtaining better clustering results. Compared with the traditional K-Means algorithm, adaptive K-Means has better clustering performance and scalability, and can handle datasets with different densities and shapes.

[0093] After clustering, k clusters are obtained. The cluster center curve Y{y1,y2,…,y} is found. k The cluster center curve is used as a typical feature library curve.

[0094] Step Ac: The method of clustering curves using the adaptive K-means algorithm includes:

[0095] Step 1: Determine the maximum number of clusters k max and minimum cluster number k min ;

[0096] Step 2: Let k = k min ;

[0097] Step 3: Cluster the power consumption curves of high-voltage users by calculating the distance between the curves using Euclidean distance based on the cluster number k;

[0098] Step 4: Calculate the evaluation index using the profile coefficient. The value ranges from -1 to 1; a larger value indicates a better effect.

[0099] Step 5: Increment k by 1, if k is not greater than k max Return to step three and continue clustering, executing steps three and four. When k = k max No further clustering, proceed to step six;

[0100] Step 6: Compare the I values ​​of each clustering result. DBI , will the maximum value (I) DBI ) max The corresponding number of cluster centers k is used as the result of clustering.

[0101] Step Ba: The method for obtaining the reference curve includes:

[0102] The current missing data's daily energy reading curve is defined as the curve to be fitted, X. The similarity between X and the cluster center curve Y in the typical feature library is calculated, and the curve with the highest similarity is used as the reference curve.

[0103] The system matches user profile data against an industry feature library to categorize curve data by industry, region, and electricity consumption level. Then, based on the date of the electricity reading curve containing missing points, it matches the season, weather, and temperature to the curve in the industry feature library.

[0104] Based on the missing data points, the daily energy consumption curve was processed. Values ​​were recorded at 15-minute intervals from 00:00 to 24:00, resulting in 97 data points per day {q1,q2,…,q…}. 97}, 96 minimum intervals {o1,o2,…,o 96}, calculate the electricity consumption increase R{r1,r2,…,r between two adjacent minimum intervals. 95};

[0105] The Z-Score is used to standardize the increase data. The Z-Score standardization method can eliminate the amplitude difference between different time points by using the daily electricity reading curve data, so that the mean is 0 and the deviation is 1, and the similarity between them can be compared more accurately.

[0106] The specific method is to first calculate the mean and standard deviation of the curve X to be fitted, and then calculate the standardized value of each data point according to the Z-Score standardization formula. The specific calculation formula is as follows:

[0107]

[0108] In the above formula, r xi ′ represents the standardized increase data, r xi This is the original increase data, μ is the average increase of the curve, and σ is the standard deviation of the increase data of the curve;

[0109] Based on standardized increase data, the curve to be fitted, X, and the cluster center, Y{y1,y2,…,y} are calculated. k The Euclidean distance of} is used to obtain the reference curve. The specific calculation formula is as follows:

[0110]

[0111] In the above formula, r xi ′ refers to the i-th standardized increase data on the curve X to be fitted, r yi ′ refers to the i-th standardized increase data on curve Y, and d(X,Y) represents the similarity between curve X and curve Y.

[0112] Compare k d(X,Y) values ​​and find the minimum Euclidean distance (d(X,Y)). min The corresponding cluster center Y′ is used as the reference curve for the curve X to be fitted.

[0113] Step Ca: The method for replacing the energy readings at the missing points includes:

[0114] Obtain the energy readings at 97 points on the reference curve Y′, and calculate the energy consumption increase {r1, r2, ... r} between 95 adjacent minimum intervals. 95}

[0115] Identify the missing points on the curve X to be fitted and label their time series. Sort the missing points according to their time series, calculate the energy readings for each missing point, and use these calculated energy readings as replacement values. The formula for calculating the energy readings for the missing points is as follows:

[0116]

[0117] In the above formula, q i The energy reading is the time series of missing points on curve X, where q is the energy reading. i-1 q i+1 It is a missing point q i The electrical energy readings at the two points before and after, r i-1 It represents the increase in the reference curve and the time series corresponding to the missing points.

[0118] Step Da: The data verification method includes:

[0119] The Grubbs criterion was used to identify outliers, and the electricity readings {q1,q2,…,q} of 97 points on the day the missing point occurred were obtained. 97 Arrange the data in descending order and count q1 <... 97 The standard deviation is calculated using the following formula:

[0120] ​

[0121] The Grubbs value is calculated using the following definition:

[0122]

[0123] In the formula, μ is q i The average value is g i For Grubbs values, σ is the average of the 97 measurements, and σ is the standard deviation of the 97 values.

[0124] With a confidence probability P = 0.95 and a critical value g... p (97) = 3.196, if g i If the value is ≥3.196, then i is an outlier and needs to be removed and replaced by other methods. For example, the energy consumption curve of the user's curve to be fitted one day ago can be used as a reference curve to replace the original fitting result; the energy consumption curve of the same period last month can also be used as a reference curve to replace the original fitting result; or the energy consumption curve of the same period last year can also be used as a reference curve to replace the original fitting result.

[0125] To illustrate the implementation of this invention in more detail, actual data will be used for explanation below.

[0126] Taking a high-voltage user within the jurisdiction of a power supply company in a certain city of a certain province as an example, this user is in the textile and garment manufacturing industry, with a voltage level of 380V. Based on the historical energy reading curve of a three-phase three-wire energy meter for this user, it is known that the user's energy readings on March 5th were missing at four points: 15, 28, 42, and 43 (see Table 1). First, the user's feature information is extracted as shown in Table 2, matched against the industry feature database, and a reference curve is found. Then, the user's energy reading curve for that day is fitted using the reference curve, resulting in 15 = 3957.80, 28 = 3963.52, 42 = 3967.86, and 43 = 3968.01. Finally, the fitted values ​​of 15, 28, 42, and 43 are validated using the Grubbs criterion for identifying outliers, and the largest g among the 97 points is calculated. i =1.72. Query the Grubbs critical value g. p (97) = 3.196, g i Since there were no values ​​greater than 3.196, it can be determined that there were no outliers. The fitted energy readings were retained, and the energy reading curve for this user on March 5th was successfully fitted.

[0127] Table 1. Energy reading curve of a user's three-phase three-wire energy meter on March 5th:

[0128]

[0129]

[0130] Table 2: User's characteristic information:

[0131] industry area voltage level season weather temperature Woven garment manufacturing A certain district of a certain city in a certain province 380V spring good cold

[0132] Example 2:

[0133] A high-voltage user curve fitting device based on a typical industry feature library, comprising:

[0134] Industry Feature Library Construction Module: Used to construct an industry feature library based on high-voltage user profile data and power consumption indicator curve data;

[0135] Reference curve acquisition module: used to acquire user data with missing power readings, classify the daily power reading curve data with missing points, calculate the similarity with the cluster center curves in the industry feature library using standardized Euclidean distance, and obtain reference curves;

[0136] Energy reading replacement module: This module is used to fit the current daily energy reading curve with the acquired reference curve and replace the missing energy readings.

[0137] Data validation module: Used to validate the replacement value and determine whether the replacement method needs to be changed based on the validation result.

[0138] The high-voltage user curve fitting device based on a typical industry feature library provided in this embodiment of the invention can execute the high-voltage user curve fitting method based on a typical industry feature library provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0139] Example 3:

[0140] This invention also provides a computer device, including a processor and a storage medium;

[0141] The storage medium is used to store instructions;

[0142] The processor is configured to operate according to the instructions to perform the steps of the following method:

[0143] An industry feature library was constructed based on the archive data of high-voltage users and the power consumption indicator curve data.

[0144] Acquire user data with missing energy readings, categorize the daily energy reading curve data with missing points, calculate the similarity with the cluster center curves in the industry feature library using standardized Euclidean distance, and obtain reference curves;

[0145] The current daily energy consumption reading curve is fitted with the acquired reference curve, and the missing energy consumption readings are replaced.

[0146] Perform data validation on the replacement values, and determine whether the replacement method needs to be changed based on the validation results.

[0147] For details of each step in this embodiment, please refer to Embodiment 1, which will not be repeated here. Since this embodiment adopts the same technical concept as Embodiment 1, it also possesses the technical effects described in Embodiment 1.

[0148] Example 4:

[0149] This invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the following method:

[0150] An industry feature library was constructed based on the archive data of high-voltage users and the power consumption indicator curve data.

[0151] Acquire user data with missing energy readings, categorize the daily energy reading curve data with missing points, calculate the similarity with the cluster center curves in the industry feature library using standardized Euclidean distance, and obtain reference curves;

[0152] The current daily energy consumption reading curve is fitted with the acquired reference curve, and the missing energy consumption readings are replaced.

[0153] Perform data validation on the replacement values, and determine whether the replacement method needs to be changed based on the validation results.

[0154] For details of each step in this embodiment, please refer to Embodiment 1, which will not be repeated here. Since this embodiment adopts the same technical concept as Embodiment 1, it also possesses the technical effects described in Embodiment 1.

[0155] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied 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.

[0156] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0157] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0158] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0159] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for high-voltage user curve fitting based on a typical industry feature library, characterized in that, include: An industry feature library was constructed based on the archive data of high-voltage users and the power consumption indicator curve data. Acquire user data with missing energy readings, categorize the daily energy reading curve data with missing points, calculate the similarity with the cluster center curves in the industry feature library using standardized Euclidean distance, and obtain reference curves; The current daily energy consumption reading curve is fitted with the acquired reference curve, and the missing energy consumption readings are replaced. Perform data validation on the replacement values, and determine whether the replacement method needs to be changed based on the validation results; The method for constructing the industry feature library includes: Acquire the profile data and power consumption curve data of high-voltage users; Based on the acquired archival data, the dimensions that affect the electricity data in the existing archival data are extracted and classified. The classified dimensions include industry, region and electricity consumption level. Obtain the daily energy consumption reading curve data for each category, and segment and classify the curve data according to the date corresponding to the curve data. The classification dimensions include season, weather and temperature. Cluster the daily electricity consumption curves; The clusters obtained from clustering are named according to the feature dimensions, and an industry feature library is constructed with industry + region + electricity consumption level + season + weather + temperature + daily energy reading curve. The clustering method includes: Data processing is performed on the daily electricity consumption indicator curve to calculate the electricity consumption increase data between two adjacent minimum intervals; The price increase data was standardized using the Z-Score standardization method. Based on the standardized increase data, the Euclidean distance formula is used to calculate the curve similarity. An adaptive K-means algorithm is used to cluster the curves.

2. The method for high-voltage user curve fitting based on a typical industry feature library according to claim 1, characterized in that, The method for clustering curves using the adaptive K-means algorithm includes: Step 1: Determine the maximum number of clusters k max and minimum cluster number k min ; Step 2: Let k = k min ; Step 3: Cluster the power consumption curves of high-voltage users by calculating the distance between the curves using Euclidean distance based on the cluster number k; Step 4: Calculate the evaluation index using the profile coefficient. The value range is [-1, 1]; Step 5: Increment k by 1, if k is not greater than k max Return to step three and continue clustering, performing steps three and four until k = k max No further clustering, proceed to step six; Step 6: Compare the I values ​​of each clustering result. DBI , will the maximum value (I) DBI ) max The corresponding number of cluster centers k is used as the result of clustering.

3. The method for high-voltage user curve fitting based on a typical industry feature library according to claim 1, characterized in that, The method for obtaining the reference curve includes: Match industry feature databases based on user profile data; Based on the missing points, process the daily energy consumption curve and calculate the electricity consumption increase data between two adjacent minimum intervals; The Z-Score was used to standardize the price increase data; Based on standardized growth rate data, the Euclidean distance between the current curve and each cluster center is calculated to obtain a reference curve; Specifically as follows: Calculate the mean and standard deviation of the curve X to be fitted, and then calculate the standardized value of each data point using the Z-Score standardization formula. The formula is as follows: In the formula, r xi ′ represents the standardized increase data, r xi This is the original increase data, μ is the average increase of the curve, and σ is the standard deviation of the increase data of the curve; Based on standardized increase data, the curve to be fitted, X, and the cluster center, Y{y1,y2,…,y} are calculated. k The Euclidean distance to} is calculated using the following formula: In the formula, r xi ′ represents the i-th standardized increase data on the curve X to be fitted, r yi ′ refers to the i-th standardized increase data on curve Y, and d(X,Y) represents the similarity between curve X and curve Y; Compare k d(X,Y) values ​​and find the minimum Euclidean distance (d(X,Y)). min The corresponding cluster center Y′ is used as the reference curve for the curve X to be fitted.

4. The method for high-voltage user curve fitting based on a typical industry feature library according to claim 3, characterized in that, The method for replacing the power readings at missing points includes: Obtain the energy readings at 97 points on the reference curve Y′, and calculate the energy consumption increase {r1, r2, ... r} between 95 adjacent minimum intervals. 95 }; Obtain the missing points on the curve X to be fitted, mark the time series of the missing points, sort the missing points according to the time series, calculate the energy readings of the missing points in turn, and use the calculated energy readings as replacement values. The formula for calculating the energy readings of the missing points is as follows: In the above formula, q i The energy reading is the time series of missing points on curve X, where q is the energy reading. i-1 q i+1 It is a missing point q i The electrical energy readings at the two points before and after, r i-1 It represents the increase in the reference curve and the time series corresponding to the missing points.

5. The method for high-voltage user curve fitting based on a typical industry feature library according to claim 1, characterized in that, The data verification method includes: The Grubbs criterion was used to identify outliers, and the energy readings {q1,q2,…,q} of 97 points on the day the missing data occurred were obtained. 97 Arrange the data in descending order and count q1 <... 97 The standard deviation is calculated using the following formula:​ The Grubbs value is calculated using the following definition: In the formula, μ is q i The average value is g i For Grubbs values, σ is the average of the 97 measurements, and σ is the standard deviation of the 97 values. With a confidence probability P = 0.95 and a critical value g... p (97) = 3.196, if g i If the value is ≥3.196, then i is an outlier and needs to be removed and replaced by other methods. These other methods include replacing the original fitted result with the average value of the missing value over the previous 3 days.

6. A high-voltage user curve fitting device based on a typical industry feature library, characterized in that, include: Industry Feature Library Construction Module: Used to construct an industry feature library based on high-voltage user profile data and power consumption indicator curve data; Reference curve acquisition module: used to acquire user data with missing power readings, classify the daily power reading curve data with missing points, calculate the similarity with the cluster center curves in the industry feature library using standardized Euclidean distance, and obtain reference curves; Energy reading replacement module: This module is used to fit the current daily energy reading curve with the acquired reference curve and replace the missing energy readings. Data validation module: Used to validate the replacement value and determine whether the replacement method needs to be changed based on the validation result; The method for constructing the industry feature library includes: Acquire the profile data and power consumption curve data of high-voltage users; Based on the acquired archival data, the dimensions that affect the electricity data in the existing archival data are extracted and classified. The classified dimensions include industry, region and electricity consumption level. Obtain the daily energy consumption reading curve data for each category, and segment and classify the curve data according to the date corresponding to the curve data. The classification dimensions include season, weather and temperature. Cluster the daily electricity consumption curves; The clusters obtained from clustering are named according to the feature dimensions, and an industry feature library is constructed with industry + region + electricity consumption level + season + weather + temperature + daily energy reading curve. The clustering method includes: Data processing is performed on the daily electricity consumption indicator curve to calculate the electricity consumption increase data between two adjacent minimum intervals; The price increase data was standardized using the Z-Score standardization method. Based on the standardized increase data, the Euclidean distance formula is used to calculate the curve similarity. An adaptive K-means algorithm is used to cluster the curves.

7. A computer device, characterized in that, Including processor and storage media; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1 to 5.

Images