Electricity utilization data processing method and device, electronic equipment and readable storage medium

Abstract

This invention provides a method, apparatus, electronic device, and readable storage medium for processing electricity consumption data. The method includes clustering users based on a first frozen electricity consumption to obtain first cluster centers for electricity consumption patterns; determining a target electricity consumption cycle based on the changing trend of the first frozen electricity consumption; clustering users based on a second frozen electricity consumption to obtain second cluster centers; and comparing the first and second cluster centers to obtain an analysis result of the user's electricity consumption habits. Thus, this invention, without excessively increasing sampling pressure, obtains a more detailed analysis of electricity consumption behavior through a broader dimensional analysis of the first frozen electricity consumption, thereby providing a more scientific and reliable theoretical support for distribution network management and adjustments.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of Internet of Things (IoT) technology, and in particular to a method, apparatus, electronic device, and readable storage medium for processing electricity data. Background Technology

[0002] The power grid only transmits electricity, not stores it. Therefore, understanding users' electricity consumption habits for grid scheduling helps improve power distribution efficiency and conserve energy. While smart meters can collect user electricity data, their sampling frequency is limited by factors such as data processing volume, hardware performance, and data transmission and storage. Insufficient data volume leads to poor processing and analysis of electricity consumption data. Summary of the Invention

[0003] This invention provides a method, apparatus, electronic device, and readable storage medium for processing electricity consumption data, in order to solve the problem of poor performance in processing and analyzing electricity consumption data.

[0004] To solve the above problems, the present invention is implemented as follows:

[0005] In a first aspect, embodiments of the present invention provide a method for processing electricity consumption data, comprising the following steps:

[0006] Users are clustered based on their first frozen electricity consumption to obtain the first cluster center of the electricity consumption pattern, wherein the first frozen electricity consumption is the frozen electricity consumption of the user in each electricity consumption cycle;

[0007] The target power consumption cycle is determined based on the trend of the change in the first frozen power volume, wherein the target power consumption cycle is the power consumption cycle corresponding to the maximum information gain of the clustering result of the first cluster center;

[0008] Users are clustered based on the second frozen electricity consumption to obtain a second cluster center, wherein the electricity consumption cycle includes multiple electricity consumption cycles, and the second frozen electricity consumption is the frozen electricity consumption of the user in the electricity consumption cycle;

[0009] The analysis results of users' electricity consumption habits are obtained by comparing the first cluster center and the second cluster center.

[0010] In some embodiments, clustering users based on their first frozen electricity consumption to obtain a first cluster center for their electricity usage patterns includes:

[0011] Electricity consumption data of each user is acquired in units of a first time period, wherein the first time period includes multiple second sub-periods, and the electricity consumption data includes the first frozen electricity in each of the second sub-periods within the first time period;

[0012] The electricity consumption data is classified according to the distance between the electricity consumption data and each central data, wherein the central data includes k electricity consumption data selected from the electricity consumption data, where k is an integer greater than 1;

[0013] Using the center of gravity of each type of electricity consumption data as the center, the center of the electricity consumption data is iteratively updated.

[0014] Under certain preset conditions, the center of the iteratively updated electricity consumption data is taken as the cluster center of the electricity consumption data. The preset conditions include the iterative update reaching a preset number of times or the objective function of the iterative update converging.

[0015] In some embodiments, classifying the electricity consumption data based on the distance between the electricity consumption data and the data from each center includes:

[0016] By selecting different k values, N classification methods for the electricity consumption data are obtained, where N is an integer greater than 1 with a preset value, and k is less than or equal to N+1.

[0017] The step of using the center of the iteratively updated electricity consumption data as the cluster center of the electricity consumption data under the condition of meeting preset conditions includes:

[0018] The classification method with the optimal objective function among the N classification methods is used as the classification method for the electricity consumption data, and the cluster centers of different types of electricity consumption data are determined. The classification method with the optimal objective function is the one whose objective function converges first or whose objective function converges to the smallest value.

[0019] In some embodiments, before determining the target electricity consumption cycle based on the trend of the change in the first frozen electricity amount, the method further includes:

[0020] The change in the first frozen power level is determined based on the user's first frozen power level.

[0021] The change is normalized to obtain the trend of the change, wherein the normalization is achieved by the following formula:

[0022]

[0023] Where D is the normalization result, D i,j σ represents the change in the first frozen electricity amount between two adjacent electricity consumption cycles, σ is the standard deviation of the change, and μ is a preset coefficient.

[0024] In some embodiments, determining the target electricity consumption cycle based on the trend of the change in the first frozen electricity amount includes:

[0025] Calculate the information entropy of the first frozen power quantity when dividing it based on different attributes;

[0026] Decision trees are constructed using the attribute with the highest information gain as the splitting criterion.

[0027] The electricity consumption cycle corresponding to the node with the most information in the decision tree is taken as the target electricity consumption cycle.

[0028] In some embodiments, clustering users based on the second frozen battery level to obtain a second cluster center includes:

[0029] The target electricity consumption cycle is divided into multiple electricity consumption cycles;

[0030] Adjust the sampling mode of the smart meter to obtain the second frozen electricity consumption of each user in each electronic cycle through the smart meter;

[0031] Users are clustered based on the second frozen charge in each electron cycle to obtain the second cluster center.

[0032] Secondly, embodiments of the present invention provide an electricity data processing device, comprising:

[0033] The first clustering module is used to cluster users based on their first frozen electricity consumption and obtain the first cluster center of the electricity consumption pattern.

[0034] The determination module is used to determine the target power consumption cycle based on the trend of the change in the first frozen power consumption, wherein the target power consumption cycle is the power consumption cycle corresponding to the maximum information gain of the clustering result of the first cluster center;

[0035] The second clustering module is used to cluster users according to the second frozen electricity to obtain the second cluster center, wherein the electricity consumption cycle includes multiple electricity consumption cycles, and the second frozen electricity is the first frozen electricity of the user in the electricity consumption cycle;

[0036] The analysis module is used to compare the first cluster center and the second cluster center to obtain the analysis results of the user's electricity consumption habits.

[0037] In some embodiments, the first clustering module includes:

[0038] The acquisition submodule is used to acquire the electricity consumption data of each user in a first time period, wherein the first time period includes multiple second sub-periods, and the electricity consumption data includes the frozen electricity in each of the second sub-periods within the first time period;

[0039] The classification submodule is used to classify the electricity consumption data according to the distance between the electricity consumption data and each central data, wherein the central data includes k electricity consumption data selected from the electricity consumption data, where k is an integer greater than 1;

[0040] The iterative submodule is used to iteratively update the center of the electricity consumption data, taking the center of gravity of each type of electricity consumption data as the center.

[0041] A determination submodule is used to determine the center of the iteratively updated electricity consumption data as the cluster center of the electricity consumption data under the condition that the preset conditions are met, wherein the preset conditions include the iterative update reaching a preset number of times or the objective function of the iterative update converging.

[0042] Thirdly, embodiments of the present invention also provide an electronic device, including: a memory, a processor, and a program stored in the memory and executable on the processor; the processor is configured to read the program in the memory to implement the steps in the method described in the first aspect above.

[0043] Fourthly, embodiments of the present invention also provide a readable storage medium for storing a program, which, when executed by a processor, implements the steps of the method described in the first aspect above.

[0044] In this embodiment of the invention, users are clustered based on their first frozen electricity consumption to obtain a first cluster center for their electricity consumption patterns. Using the maximum information gain of the first frozen electricity consumption as a dividing criterion, a target electricity consumption cycle is determined based on the trend of changes in the first frozen electricity consumption. Users are then clustered based on the first frozen electricity consumption of multiple electricity consumption cycles included in the target electricity consumption cycle to obtain a second cluster center. The first and second cluster centers are compared to obtain the user's electricity consumption habit analysis results. Thus, by introducing the analysis and interpretation of the first frozen electricity consumption cluster centers, a scientific basis is provided for adjusting the cluster centers and the sampling of the first frozen electricity consumption. Without excessively increasing the sampling pressure, a more detailed analysis of residential electricity consumption behavior is obtained through a broader dimensional analysis of the first frozen electricity consumption, thereby providing a more scientific and reliable theoretical support for distribution network management adjustments. Attached Figure Description

[0045] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a flowchart illustrating an embodiment of the electricity data processing method provided by the present invention;

[0047] Figure 2 This is a schematic diagram illustrating the trend of change in the first frozen electricity volume according to an embodiment of the present invention;

[0048] Figure 3 This is a partial schematic diagram of a decision tree according to an embodiment of the present invention;

[0049] Figure 4 This is a schematic diagram of user distribution in one embodiment of the present invention;

[0050] Figure 5 This is a schematic diagram of the structure of an electricity data processing device provided in an embodiment of the present invention;

[0051] Figure 6 This is a schematic diagram of the structure of an electronic device provided in one embodiment of the present invention. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] In the embodiments of this invention, the terms "first," "second," etc., are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices. Additionally, the use of "and / or" in this application indicates at least one of the connected objects, such as A and / or B and / or C, representing seven possibilities: including A alone, B alone, C alone, and the presence of both A and B, both B and C, both A and C, and the presence of A, B, and C.

[0054] This invention provides a method for processing electricity consumption data.

[0055] like Figure 1 As shown, in one embodiment, the electricity consumption data processing method includes the following steps:

[0056] Step 101: Cluster users based on their first frozen electricity consumption to obtain the first cluster center of their electricity consumption patterns.

[0057] In this embodiment, the first frozen electricity consumption refers to the user's frozen electricity consumption in each electricity consumption cycle. Specifically, it can refer to the user's electricity consumption per unit time. It is understood that currently, user electricity consumption can be collected through smart meters. However, due to limitations in data transmission and storage conditions, the collection and transmission of user electricity consumption data cannot be too frequent. Therefore, in this embodiment, the user's first frozen electricity consumption is collected according to a set electricity consumption cycle. For example, an electricity consumption cycle can be a day or an hour.

[0058] It should be noted that the acquisition, storage, and application of user personal information involved in the technical solution of this embodiment all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0059] After obtaining the first frozen power consumption of multiple users, further clustering mining is performed on the first frozen power consumption to obtain multiple first cluster centers, where each cluster center corresponds to a power consumption mode.

[0060] Here, electricity usage patterns can be understood as different electricity consumption habits of users. It's understandable that different groups of people have different electricity consumption habits. For example, retirees may be accustomed to watching TV during the day and going to bed early in the evening; therefore, their daytime electricity consumption is higher than their nighttime consumption. People who need to work during the day may watch TV or use electronic devices for entertainment in the evening, and since they are not at home during the day, they do not use electrical appliances; therefore, their daytime electricity consumption is lower than their nighttime consumption.

[0061] In this way, different groups of people have different electricity consumption habits due to differences in their living habits and lifestyles. The first frozen electricity consumption can directly reflect their electricity consumption situation. Thus, cluster analysis can be performed based on the first frozen electricity consumption to obtain their electricity consumption patterns.

[0062] In some embodiments, step 101 specifically includes:

[0063] Electricity consumption data of each user is acquired in units of a first time period, wherein the first time period includes multiple second sub-periods, and the electricity consumption data includes the first frozen electricity in each of the second sub-periods within the first time period;

[0064] The electricity consumption data is classified according to the distance between the electricity consumption data and each central data, wherein the central data includes k electricity consumption data selected from the electricity consumption data, where k is an integer greater than 1;

[0065] Using the center of gravity of each type of electricity consumption data as the center, the center of the electricity consumption data is iteratively updated.

[0066] Under certain preset conditions, the center of the iteratively updated electricity consumption data is taken as the cluster center of the electricity consumption data. The preset conditions include the iterative update reaching a preset number of times or the objective function of the iterative update converging.

[0067] For example, in this embodiment, "day" is used as the first time period. Further, the first time period includes multiple second time periods "hour". In this embodiment, the user's electricity consumption data for multiple days is obtained, wherein the electricity consumption data for each day is composed of the first frozen electricity consumption of each hour in the 24 hours of that day.

[0068] Next, these electricity consumption data are categorized. In this embodiment, k electricity consumption data points are first selected as the core data. The value of k can be an integer greater than 1, such as 2, 3, 4, etc., and its specific value range can be set based on experience. Here, the selected k electricity consumption data points can be arbitrarily chosen. For example, they can be selected by generating random numbers, directly specified, or extracted according to certain rules.

[0069] Next, the distance between the electricity consumption data for each user and the central data is calculated, and the electricity consumption data is assigned to the data type corresponding to the central data closest to it. This yields k types of electricity consumption data. Then, for each type of electricity consumption data, the centroid of each type is calculated and used as the new center of that category. This process is iteratively executed, updating the position of the data centers. When certain conditions are met, such as a certain number of iterations or when the objective function converges, the center determined at this point is taken as the cluster center for that type; in this embodiment, it is denoted as the first cluster center.

[0070] The specific steps for classifying the electricity consumption data based on its distance from the data of each center include:

[0071] By selecting different k values, N classification methods for the electricity consumption data are obtained, where N is an integer greater than 1 with a preset value, and k is less than or equal to N+1.

[0072] The step of using the center of the iteratively updated electricity consumption data as the cluster center of the electricity consumption data under the condition of meeting preset conditions includes:

[0073] The classification method with the optimal objective function among the N classification methods is used as the classification method for the electricity consumption data, and the cluster centers of different types of electricity consumption data are determined.

[0074] It should be understood that if the value of k is equal to 1, it means that the electricity consumption habits of each user are basically the same. However, the electricity consumption habits of different users are usually more complex. Therefore, it is meaningless to study their electricity consumption habits as a type. Thus, in this embodiment, the value of k is greater than or equal to 2. Specifically, following the above steps, the value of k is taken as an integer greater than 2, such as 2, 3, 4, etc., and the classification results are obtained under different values ​​of k.

[0075] For the multiple classification results obtained, this embodiment further selects the optimal classification result. For example, the optimal condition can be that the objective function converges first, or that the objective function converges to the smallest value. After determining the optimal classification result, the cluster center corresponding to this classification result is taken as the first cluster center.

[0076] It should be noted that the calculation method for calculating the cluster centers of data can be found in relevant technologies, and will not be elaborated here.

[0077] like Figure 2 As shown, in one embodiment, after comparison, it is found that the classification result corresponding to k equal to 10 is the optimal clustering result. Then, the user's electricity consumption data is divided into 10 different types of electricity consumption patterns. Based on the clustering results, the typical values ​​of the electricity consumption characteristics of these 10 electricity consumption patterns can be obtained respectively. For example, the electricity consumption data corresponding to the center of each type of electricity consumption pattern can be used as the typical value of the electricity consumption characteristics of that electricity consumption pattern.

[0078] Step 102: Determine the target electricity consumption cycle based on the trend of the change in the first frozen electricity amount;

[0079] Next, the obtained cluster centers will be further analyzed and adjusted.

[0080] In this embodiment, the target electricity consumption cycle is the electricity consumption cycle corresponding to the maximum information gain of the clustering results of the first cluster center. During implementation, the change in the user's first frozen electricity consumption is first obtained, and then the trend of this change is obtained based on this change.

[0081] The change in the first frozen electricity amount D i,j It can be obtained through the following formula:

[0082]

[0083] Among them, E i,j Let D represent the first frozen electricity in the j-th hour of the i-th day. Thus, when j=0, the change is D. i,j The first frozen electricity E from 0:00 to 1:00 on day i. i,0 The first frozen electricity E from 23:00 to 24:00 on day i-1 i-1,23The difference, when j is an integer from 1 to 23, is expressed as a change in D. i,j It is the difference in the first frozen power of two consecutive hours.

[0084] Generally, the obtained change data roughly conforms to a normal distribution. Under the function curve, 68% of the area is within one standard deviation of the mean, and 99% of the area is within four standard deviations of the mean.

[0085] It should be understood that the change is a specific value, so the data obtained is difficult to reflect the degree of increase or decrease in the change.

[0086] For example, if a user consumes 1 kWh of electricity in the first hour and 2 kWh in the next hour, the electricity consumption increases by 100%, and the change is 1 kWh. For another user, if they consume 5 kWh in the first hour and 6 kWh in the next hour, the electricity consumption increases by 20%, and the change is also 1 kWh. Therefore, the change only reflects the magnitude of the change in a user's electricity consumption, but it cannot accurately reflect the changes in a user's electricity usage.

[0087] Therefore, this embodiment further normalizes the change in the first frozen charge. Before step 102, it also includes:

[0088] The change in the first frozen power level is determined based on the user's first frozen power level.

[0089] The change is normalized to obtain the trend of the change, wherein the normalization is achieved by the following formula:

[0090]

[0091] Where D is the normalization result, D i,j Let σ be the change in the first frozen electricity consumption over two adjacent electricity consumption cycles, σ be the standard deviation of the change, and μ be a preset coefficient. The sign of the normalized result D and the change D are related. i,j Consistent. Thus, after normalization, the trend of user electricity consumption changes can be displayed in a form that is close to 0 and fluctuates around 0. When a user's electricity load increases, the normalization result is positive; when electricity consumption decreases, the normalization result is negative; if the electricity consumption remains constant, the normalization result is 0. The absolute value of the normalization result represents different degrees of change; obviously, the larger the absolute value of the normalization result, the more drastic the change.

[0092] This normalization process allows the clustering center results of the first frozen charge difference to be mapped to a finite-dimensional space.

[0093] In this embodiment, before step 102, the change in the first frozen electricity amount in two adjacent electricity consumption cycles can be determined by formula (1), and then normalization is performed by formula (2). The result of the normalization process can reflect the changing trend of the first frozen electricity amount.

[0094] After normalizing the first cluster centers, we further analyze the obtained first cluster centers to extract the more core information contained therein.

[0095] It is important to note that, based on the obtained cluster centers, we are more interested in finding the most obvious differences between different cluster centers, so as to find the reasons for the differences between different first-freeze power cluster centers.

[0096] In this embodiment, the target electricity consumption cycle is the electricity consumption cycle corresponding to the maximum information gain of the clustering results of the first cluster center. This can be understood as the target electricity consumption cycle having the highest information gain among multiple electricity consumption cycles, or, in other words, determining the target electricity consumption cycle among multiple electricity consumption cycles based on the maximum information gain as the partitioning condition. Using the attribute with the highest information gain as a reference allows for the partitioning of results with less information, even when the sampling frequency of the smart meter is limited, thus obtaining the processing results for the electricity consumption data.

[0097] Different attributes have a significant impact on electricity consumption. For example, in some regions, electricity prices are relatively low late at night, so changes in electricity prices can be an attribute that may affect electricity consumption. Similarly, in a certain area, most companies start work at 8 a.m., and users typically leave home for work around 7 a.m. After leaving home, users' electricity consumption decreases, so work hours may also be an attribute that affects electricity consumption. Temperature may affect users' use of air conditioning, so weather factors can also be a data attribute that affects electricity consumption.

[0098] In this embodiment, a decision tree can be constructed based on electricity consumption data, and the target electricity consumption cycle can be further determined based on the established decision tree.

[0099] In one embodiment, step 103 specifically includes:

[0100] Calculate the information entropy of the first frozen power quantity when dividing it based on different attributes;

[0101] Decision trees are constructed using the attribute with the highest information gain as the splitting criterion.

[0102] The electricity consumption cycle corresponding to the node with the most information in the decision tree is taken as the target electricity consumption cycle.

[0103] In one embodiment, constructing a decision tree can be understood as follows: First, given a dataset T containing t samples, each sample belongs to one of m different categories Ci (i = 1, 2, 3, ..., m), where the number of samples in category Ci is ti, and Pi is the probability that a sample belongs to Ci, taking Pi = ti / t, then the expected amount of information required to classify a given sample is:

[0104]

[0105] Let Q be an attribute in set T, and Q have n distinct values ​​{q1, q2, ..., qn}. n Q can divide T into n distinct classes {T1, T2, ..., T}. n}, T i The number of samples is t 1j +t 2j +…...+t mj Then T i The probability p that belongs to Ci ij =t ij / (t 1j +t 2j +…...+t mj Thus, for T i The expected information is:

[0106]

[0107] Thus, the information entropy E(Q) of the training sample set partitioned by Q is:

[0108]

[0109] Thus, the information entropy E(Q) of dataset T can be obtained.

[0110] In this way, by using the maximum information gain of the first frozen power as the dividing condition, different information entropy E(Q) of the first frozen power is obtained by setting different information Q values. The larger the value of E, the more information that the attribute Q contributes to the cluster center. Thus, the attribute Q with the maximum information gain of each node is the key attribute for establishing the decision tree of the first frozen power.

[0111] In the above formula, the user's electricity consumption attribute Q needs to be extracted based on the user's electricity consumption data. In one embodiment, it can be understood that because different users have different electricity consumption habits, their electricity consumption data at different times is different. Specifically, it may be reflected in the magnitude of characteristic points such as peak power consumption, trough power consumption, peak-to-trough power difference, and total load. It may also include the time when these characteristic points appear and the changing trend of these characteristics.

[0112] These characteristics can all reflect a user's electricity consumption behavior. For example, peak power consumption may indicate that all or most of the appliances in the user's home are working. Further, it can be determined that the user may be cooking with appliances. Combining the peak value with the peak value, the number of appliances that are working can be estimated. On the other hand, trough power consumption corresponds to some or all appliances stopping working, which may indicate that the user is away from home. Combining the trough value with the trough value, it can be inferred whether there are appliances such as refrigerators that are left on standby for a long time, and the estimated number of standby appliances. Furthermore, the timing of these characteristics can also reflect the user's lifestyle and daily routines to some extent, such as daily working hours, cooking time, and whether they work on weekends.

[0113] Information entropy E(Q) is the amount of information included in the electricity consumption data, which can also be understood as the orderliness and purity of the data. In this embodiment, it can be specifically understood as the degree of similarity between the electricity consumption data and specific data features, based on information entropy.

[0114] like Figure 3 As shown, a decision tree for electricity consumption data can be constructed based on the attribute Q, which has the maximum information gain for each node.

[0115] In this embodiment, 10:00 in the decision tree represents the data at 10 o'clock, and 10 o'clock corresponds to the first node. 1:00 represents the second node, corresponding to the electricity consumption data at 1 o'clock, which is a second-level node. Similarly, 8:00 represents the electricity consumption data at 8 o'clock, which corresponds to a third-level node. 21:00 represents the data at 9 o'clock, which corresponds to another third-level node.

[0116] The data on the right side of each equal sign represents the characteristic values ​​of different types of electricity consumption data. For example, data in similar formats such as -2:1 (1.0) and -3:5 (1.0) represent the characteristic values ​​of different types of electricity consumption data.

[0117] This decision tree can be understood as grouping all electricity consumption data with a certain electricity consumption characteristic at 10 o'clock into one category. The first row of the classification corresponds to the feature value -2, and the second to sixth rows correspond to the electricity consumption data with the feature value -1. The electricity consumption data with the feature value -1 can be further classified according to its electricity consumption characteristic at 1 o'clock, and so on, which can generate a decision tree.

[0118] The point with the largest amount of frozen electricity information is used as the key analysis node for electricity consumption analysis, and this key analysis node corresponds to the selected target electricity consumption cycle.

[0119] Step 103: Cluster users based on the second frozen power level to obtain the second cluster center.

[0120] In this embodiment, the power consumption cycle includes multiple power consumption cycles, and the second frozen power consumption is the frozen power consumption of the user during the power consumption cycle.

[0121] In some embodiments, step 103 specifically includes:

[0122] The target electricity consumption cycle is divided into multiple electricity consumption cycles;

[0123] Adjust the sampling mode of the smart meter to obtain the second frozen electricity consumption of each user in each electronic cycle through the smart meter;

[0124] Users are clustered based on the second frozen charge in each electron cycle to obtain the second cluster center.

[0125] In order to further analyze users' electricity consumption habits during the core period and obtain electricity consumption data for that core period.

[0126] In implementation, this embodiment divides the target electricity consumption cycle into multiple electricity consumption cycles. For example, if the target electricity consumption cycle is one hour, then the electricity consumption cycle is set to 5 minutes. In this way, the second frozen electricity consumption is obtained every 5 minutes within this one-hour target electricity consumption cycle. The data acquisition is specifically achieved by adjusting the sampling mode of the smart meter.

[0127] It should be understood that smart meters have certain limitations in data storage and transmission. This embodiment only processes the target electricity consumption cycle, which can collect as much effective data as possible without excessively increasing the load on the smart meter, thereby improving the effectiveness of the analysis results.

[0128] After obtaining the second frozen power, the second frozen power is further clustered by referring to the classification process in step 101 above, and the corresponding second cluster centers are obtained.

[0129] In this embodiment, five second cluster centers are used as an example.

[0130] Step 104: Compare the first cluster center and the second cluster center to obtain the user's electricity consumption habit analysis results.

[0131] Based on the first and second cluster centers obtained, further analysis can be conducted to obtain users' electricity consumption habits.

[0132] For example, in a set of data, the user types corresponding to the second cluster center include five types: a, b, c, d, and e. The first cluster center includes ten types, numbered 0 to 9. Cross-comparison reveals that types a, b, and c account for 72%, while types 0, 2, and 9 account for 85%. Comparing the user sources of types a, b, and d with those of types 0, 2, and 9 shows an overlap rate of less than 83%. The same user group has similar, unchanging electricity consumption behavior, but the results differ significantly when clustering is done 24 hours with a focus on 2 hours. This may indicate that a large number of residents in this area go to work in the morning, and these residents have similar electricity consumption habits when away from home. This leads to a decrease in the number of cluster centers in cluster 3, while the cluster overlap between the first and second cluster centers is very low. User sources of type A mainly correspond to 2 and 9, user sources of type B mainly correspond to 2 and 3, and user sources of type C mainly correspond to 6 and 9. Under these circumstances, based on the typical values ​​of the key 2-hour freeze and 24-hour first freeze electricity cluster centers, we can deduce the comprehensive household electricity consumption habits of families with high-power appliances (such as electric water heaters) that are still turned on when they are away from home.

[0133] For example, such as Figure 4 As shown in this embodiment, through analysis, it is possible to determine the distribution ratio of users who have the habit of leaving high-power electrical appliances on even when away from home in different communities within a specific area.

[0134] Clearly, further analysis based on the data from the first and second cluster centers can yield other conclusions, which will help to analyze users' electricity consumption habits more deeply, so as to rationally regulate the power supply of the grid based on the results, improve the accuracy of grid regulation, and save energy.

[0135] The technical solution in this embodiment further introduces the analysis and interpretation of cluster centers from two clustering processes of frozen electricity, providing a scientific basis for adjusting cluster centers and frozen electricity sampling. Without excessively increasing sampling pressure, a more detailed analysis of residential electricity consumption and lifestyle behavior is obtained through a broader dimensional analysis of frozen electricity, thus providing more scientific and reliable theoretical support for distribution network management adjustments.

[0136] In one embodiment, a first frozen electricity consumption is first collected using a smart meter as the user's electricity consumption data. For example, sampling can be performed once every hour. Then, the collected electricity consumption data is analyzed to obtain ten first cluster centers. After further analysis, the target electricity consumption cycle is determined to be from 8 to 10 o'clock. Next, a second frozen electricity consumption from 8 to 10 o'clock is collected using the smart meter. For example, sampling can be performed once every 10 minutes. Then, the second frozen electricity consumption from 8 to 10 o'clock is clustered to obtain five second cluster centers.

[0137] Next, comparing the first and second cluster centers, the analysis revealed that the top four user groups corresponding to the ten first cluster centers accounted for approximately 36.9%, 35.6%, 12.9%, and 5.5% of the total, respectively, while the user groups corresponding to the other first cluster centers were all less than 5%. In contrast, the user groups corresponding to the second cluster centers accounted for approximately 21.3%, 16.3%, 9.8%, and 3% of the total, respectively. This significant difference in user groups between the first and second cluster centers indicates that the user electricity consumption data contains much more information to be discovered.

[0138] For example, analyzing electricity consumption data, the peak consumption of the first cluster center, corresponding to the largest number of users (36.9%), is approximately 0.2 to 0.4 kWh; the peak consumption of the first cluster center, corresponding to the second largest number of users (35.6%), is approximately 0.05 to 0.2 kWh; the electricity consumption of the second cluster center, corresponding to the largest number of users (49.3%), is approximately 0.18 to 0.2 kWh; and the peak consumption of the second cluster center, corresponding to the second largest number of users (21.3%), is approximately 0.35 to 0.37 kWh. It should be noted that the electricity consumption corresponding to the first cluster center is the first frozen electricity consumption over one hour, while the electricity consumption corresponding to the second cluster center is the second frozen electricity consumption over 10 minutes.

[0139] This shows that users' electricity consumption is relatively concentrated between 8 and 10 o'clock, and many users consume relatively more electricity within a unit of time. This indicates that users used high-power appliances in a short period of time. Based on the specific time, it can be inferred that these were high-power cooking devices such as microwave ovens, induction cookers, and rice cookers.

[0140] Assuming that the total electricity required for cooking a regular breakfast is about 0.2 to 0.4 kWh, it can be assumed that most users cook breakfast between 8 and 10 a.m. By further analyzing the second cluster center, the specific time period during which users cook can be determined, thus allowing us to infer the users' specific electricity consumption habits.

[0141] In this way, by deeply exploring the differences in the proportion of users corresponding to the first and second cluster centers, we can obtain more analysis of users' electricity consumption habits and behaviors, thereby making targeted adjustments to power supply and grid regulation, which helps to provide more scientific and reliable theoretical support for distribution network management adjustments.

[0142] The present invention also provides an electrical data processing device.

[0143] like Figure 5 As shown, in one embodiment, the power consumption data processing device 500 includes:

[0144] The first clustering module 501 is used to cluster users based on their first frozen electricity consumption to obtain the first cluster center of the electricity consumption pattern.

[0145] The determining module 502 is used to determine the target power consumption cycle based on the trend of the change in the first frozen power amount, wherein the target power consumption cycle is the power consumption cycle corresponding to the maximum information gain of the clustering result of the first cluster center;

[0146] The second clustering module 503 is used to cluster users according to the second frozen electricity to obtain the second cluster center, wherein the electricity consumption cycle includes multiple electricity consumption cycles, and the second frozen electricity is the frozen electricity of the user in the electricity consumption cycle;

[0147] Analysis module 504 is used to compare the first cluster center and the second cluster center to obtain the user's electricity consumption habit analysis results.

[0148] In some embodiments, the first clustering module 501 includes:

[0149] The acquisition submodule is used to acquire the electricity consumption data of each user in a first time period, wherein the first time period includes multiple second sub-periods, and the electricity consumption data includes the first frozen electricity in each of the second sub-periods within the first time period;

[0150] The classification submodule is used to classify the electricity consumption data according to the distance between the electricity consumption data and each central data, wherein the central data includes k electricity consumption data selected from the electricity consumption data, where k is an integer greater than 1;

[0151] The iterative submodule is used to iteratively update the center of the electricity consumption data, taking the center of gravity of each type of electricity consumption data as the center.

[0152] A determination submodule is used to determine the center of the iteratively updated electricity consumption data as the cluster center of the electricity consumption data under the condition that the preset conditions are met, wherein the preset conditions include the iterative update reaching a preset number of times or the objective function of the iterative update converging.

[0153] In some embodiments, the classification submodule is specifically used to: select different k values ​​respectively to obtain N classification methods for the electricity consumption data classification, wherein N is an integer greater than 1 with a preset value, and k is less than or equal to N+1;

[0154] The determination submodule is specifically used to: classify the electricity consumption data according to the classification method with the optimal objective function among the N classification methods and determine the cluster centers of different types of electricity consumption data, wherein the classification method with the optimal objective function is the classification method with the first convergence of the objective function or the classification method with the smallest convergence value of the objective function.

[0155] In some embodiments, it also includes:

[0156] The change acquisition module is used to determine the change in the first frozen power based on the user's first frozen power.

[0157] The normalization module is used to normalize the change amount to obtain the trend of the change amount, wherein the normalization is achieved by the following formula:

[0158]

[0159] Where D is the normalization result, D i,j σ represents the change in the first frozen electricity amount between two adjacent electricity consumption cycles, σ is the standard deviation of the change, and μ is a preset coefficient.

[0160] In some embodiments, the determining module 502 includes:

[0161] The information entropy calculation submodule is used to calculate the information entropy of the first frozen power when divided based on different attributes;

[0162] The decision tree construction submodule is used to construct a decision tree based on the attribute with the highest information gain as the splitting condition.

[0163] The target electricity consumption cycle determination submodule is used to take the electricity consumption cycle corresponding to the node with the most information in the decision tree as the target electricity consumption cycle.

[0164] In some embodiments, the second clustering module 503 includes:

[0165] The sub-module is used to divide the target electricity consumption cycle into multiple electricity consumption cycles;

[0166] An adjustment submodule is used to adjust the sampling mode of the smart meter in order to obtain the second frozen electricity consumption of each user in each electricity consumption cycle through the smart meter;

[0167] Users are clustered based on the second frozen charge in each electron cycle to obtain the second cluster center.

[0168] The virtual training management device 400 of this embodiment can implement all the steps of the above-described electricity data processing method embodiment and achieve essentially the same technical effect, which will not be repeated here.

[0169] This invention also provides an electronic device. Please refer to [link to relevant documentation]. Figure 6 The electronic device may include a processor 601, a memory 602, and a program 6021 stored in the memory 602 and capable of running on the processor 601.

[0170] When the electronic device is a terminal, program 6021 can be executed by processor 601 to achieve the following: Figure 1 Any steps in the corresponding method embodiments and the achievement of the same beneficial effects will not be repeated here.

[0171] When the electronic device is a network-side device, program 6021 can be executed by processor 601 to achieve the following: Figure 6 Any steps in the corresponding method embodiments and the achievement of the same beneficial effects will not be repeated here.

[0172] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by hardware related to program instructions, and the program can be stored in a readable medium.

[0173] This invention also provides a readable storage medium storing a computer program, which, when executed by a processor, can perform the above-described functions. Figure 1 Any step in the corresponding method embodiment can achieve the same technical effect, and will not be repeated here to avoid repetition.

[0174] The storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0175] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the 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 of electric power utilization data processing, characterized by, Includes the following steps: Users are clustered based on their first frozen electricity consumption to obtain the first cluster center of the electricity consumption pattern, wherein the first frozen electricity consumption is the frozen electricity consumption of the user in each electricity consumption cycle; The target power consumption cycle is determined based on the trend of the change in the first frozen power volume, wherein the target power consumption cycle is the power consumption cycle corresponding to the maximum information gain of the clustering result of the first cluster center; Users are clustered based on the second frozen electricity consumption to obtain a second cluster center, wherein the electricity consumption cycle includes multiple electricity consumption cycles, and the second frozen electricity consumption is the frozen electricity consumption of the user in the electricity consumption cycle; The analysis results of users' electricity consumption habits are obtained by comparing the first cluster center and the second cluster center; The step of clustering users based on the second frozen power level to obtain the second cluster center includes: The target electricity consumption cycle is divided into multiple electricity consumption cycles; Adjust the sampling mode of the smart meter to obtain the second frozen electricity consumption of each user in each electronic cycle through the smart meter; Users are clustered based on the second frozen charge in each electron cycle to obtain the second cluster center.

2. The method of claim 1, wherein, The step of clustering users based on their first frozen electricity consumption to obtain the first cluster center of their electricity consumption patterns includes: Electricity consumption data of each user is acquired in units of a first time period, wherein the first time period includes multiple second sub-periods, and the electricity consumption data includes the first frozen electricity in each of the second sub-periods within the first time period; The electricity consumption data is classified according to the distance between the electricity consumption data and each central data, wherein the central data includes k electricity consumption data selected from the electricity consumption data, where k is an integer greater than 1; Using the center of gravity of each type of electricity consumption data as the center, the center of the electricity consumption data is iteratively updated. Under certain preset conditions, the center of the iteratively updated electricity consumption data is taken as the cluster center of the electricity consumption data. The preset conditions include the iterative update reaching a preset number of times or the objective function of the iterative update converging.

3. The method of claim 2, wherein, The classification of electricity consumption data based on the distance between the electricity consumption data and the data from each center includes: By selecting different k values, N classification methods for the electricity consumption data are obtained, where N is an integer greater than 1 with a preset value, and k is less than or equal to N+1. The step of using the center of the iteratively updated electricity consumption data as the cluster center of the electricity consumption data under the condition of meeting preset conditions includes: The classification method with the optimal objective function among the N classification methods is used as the classification method for the electricity consumption data, and the cluster centers of different types of electricity consumption data are determined. The classification method with the optimal objective function is the one whose objective function converges first or whose objective function converges to the smallest value.

4. The method of claim 1, wherein, Before determining the target electricity consumption cycle based on the trend of the change in the first frozen electricity, the method further includes: The change in the first frozen power level is determined based on the user's first frozen power level. The change is normalized to obtain the trend of the change, wherein the normalization is achieved by the following formula: ； Where D is the normalization result, This represents the change in the first frozen electricity consumption over two adjacent electricity consumption cycles. The standard deviation of the change. These are preset coefficients.

5. The method of claim 4, wherein, The step of determining the target electricity consumption cycle based on the trend of the change in the first frozen electricity amount includes: Calculate the information entropy of the first frozen power quantity when dividing it based on different attributes; Decision trees are constructed using the attribute with the highest information gain as the splitting criterion. The electricity consumption cycle corresponding to the node with the most information in the decision tree is taken as the target electricity consumption cycle.

6. An electric data processing device, characterized by include: The first clustering module is used to cluster users based on their first frozen electricity consumption and obtain the first cluster center of the electricity consumption pattern. The determination module is used to determine the target power consumption cycle based on the trend of the change in the first frozen power consumption, wherein the target power consumption cycle is the power consumption cycle corresponding to the maximum information gain of the clustering result of the first cluster center; The second clustering module is used to cluster users according to the second frozen electricity to obtain the second cluster center, wherein the electricity consumption cycle includes multiple electricity consumption cycles, and the second frozen electricity is the frozen electricity of the user in the electricity consumption cycle; The analysis module is used to compare the first cluster center and the second cluster center to obtain the user's electricity consumption habit analysis results; The second clustering module includes: The sub-module is used to divide the target electricity consumption cycle into multiple electricity consumption cycles; An adjustment submodule is used to adjust the sampling mode of the smart meter in order to obtain the second frozen electricity consumption of each user in each electricity consumption cycle through the smart meter; Users are clustered based on the second frozen charge in each electron cycle to obtain the second cluster center.

7. The apparatus of claim 6, wherein, The first clustering module includes: The acquisition submodule is used to acquire the electricity consumption data of each user in a first time period, wherein the first time period includes multiple second sub-periods, and the electricity consumption data includes the first frozen electricity in each of the second sub-periods within the first time period; The classification submodule is used to classify the electricity consumption data according to the distance between the electricity consumption data and each central data, wherein the central data includes k electricity consumption data selected from the electricity consumption data, where k is an integer greater than 1; The iterative submodule is used to iteratively update the center of the electricity consumption data, taking the center of gravity of each type of electricity consumption data as the center. A determination submodule is used to determine the center of the iteratively updated electricity consumption data as the cluster center of the electricity consumption data under the condition that the preset conditions are met, wherein the preset conditions include the iterative update reaching a preset number of times or the objective function of the iterative update converging.

8. An electronic device comprising: A transceiver, a memory, a processor, and a program stored in the memory and executable on the processor; characterized in that the processor is configured to read the program in the memory to implement the steps of the power consumption data processing method as described in any one of claims 1 to 5.

9. A readable storage medium for storing a program, characterized in that, When the program is executed by the processor, it implements the steps in the electricity data processing method as described in any one of claims 1 to 5.

Images