Industrial and commercial user electricity consumption feature analysis method based on momentum feature model

Abstract

The present application relates to the field of data processing, in particular to a commercial and industrial user electricity feature analysis method based on momentum feature model, comprising: obtaining an electricity influence factor traceability list of commercial and industrial users based on each electricity influence factor and short-period momentum factor of electricity data; determining a traceability influence factor of each electricity influence factor on the next electricity influence factor; fusing the electricity influence factors in the electricity traceability factor list of all commercial and industrial users based on the traceability influence factor, obtaining a fusion traceability table and a fusion user of each commercial and industrial user, determining a traceability scale according to the position of the fusion factors contained in the fusion user in the fusion traceability table, and constructing a data wide table according to the traceability scale. The present application fuses the electricity influence factors of commercial and industrial users, and ensures that the data wide table is reliable and traceable.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically to a method for analyzing the electricity consumption characteristics of industrial and commercial users based on a momentum characteristic model. Background Technology

[0002] Industrial and commercial electricity consumption differs from residential electricity consumption. Industrial and commercial electricity consumption is large, and the requirements for power stability are high. To ensure orderly production and operation in industry and commerce, it is necessary to monitor electricity consumption and allocate power in a timely manner. Power grid operators often use wide data tables for industrial and commercial users to monitor their electricity consumption. These wide data tables are a data format that may be required for input in artificial intelligence and machine learning fields, allowing all features of each entity in the wide data table to be considered simultaneously during model training. This includes electricity consumption data for each industrial and commercial user, as well as attribute information such as company size and equipment operating status. Power grid operators allocate power to each industrial and commercial user based on the attributes in the wide data table and historical electricity consumption data to maintain grid stability.

[0003] However, many factors influence the electricity consumption data of industrial and commercial users, such as their regulatory policies, business operations, distribution node operation, and production equipment operating time. Not only are these influencing factors different for different industrial and commercial users, but they also interact and have a causal relationship. For example, a user's business operations directly affect the operating time of their production equipment, thus impacting their electricity consumption; similarly, the operating time of their production equipment affects the operation of distribution nodes, which in turn affects the electricity consumption of other industrial and commercial users. Therefore, simply recording industrial and commercial users' electricity consumption data and enterprise attributes in a wide data table leads to a lack of data integration between different users, making it impossible to quickly identify the root causes affecting their electricity consumption and consequently hindering rapid, accurate, and targeted power allocation. Summary of the Invention

[0004] To address the aforementioned problems, this invention provides a method for analyzing the electricity consumption characteristics of industrial and commercial users based on a momentum characteristic model.

[0005] The method for analyzing the electricity consumption characteristics of industrial and commercial users based on the momentum characteristic model of the present invention adopts the following technical solution:

[0006] This invention provides a method for analyzing the electricity consumption characteristics of industrial and commercial users based on a momentum characteristic model. The method includes the following steps:

[0007] Collect electricity consumption data from industrial and commercial users and obtain all electricity consumption influencing factors for each industrial and commercial user;

[0008] Based on the difference in short-period momentum factors between each electricity consumption influencing factor and electricity consumption data, the electricity consumption influencing factors are ranked to obtain a source list of electricity consumption influencing factors for each industrial and commercial user.

[0009] In the source tracing list of electricity consumption influencing factors for each industrial and commercial user, the source tracing influence factor of each electricity consumption influencing factor on adjacent electricity consumption influencing factors is obtained based on the difference in the time corresponding to the maximum value of adjacent electricity consumption influencing factors.

[0010] In the list of electricity consumption traceability factors for all industrial and commercial users, the electricity consumption influencing factors in the list of electricity consumption traceability factors for all industrial and commercial users are merged based on the traceability influencing factors to obtain a merged traceability table and a merged user for each industrial and commercial user, wherein the merged user contains the merged factors;

[0011] The tracing scale is determined based on the position of the fusion factors included in the fusion user in the fusion tracing table, and a wide data table is constructed based on the tracing scale.

[0012] Preferably, the specific steps for obtaining all electricity consumption influencing factors for each industrial and commercial user are as follows:

[0013] The attribute values ​​of the same attribute for each industrial and commercial user at different times constitute an attribute sequence, and the electricity consumption data of each industrial and commercial user at different times constitute an electricity consumption data sequence. The grey relational analysis method is used to obtain the weight of each attribute based on all the attribute sequences and electricity consumption data sequences of each industrial and commercial user. Attributes with weights greater than a preset first threshold are recorded as electricity consumption influencing factors for each industrial and commercial user.

[0014] Preferably, the specific steps for obtaining the short-period momentum factor are as follows:

[0015] For each electricity consumption influencing factor of each industrial and commercial user, the value of each electricity consumption influencing factor at different times constitutes an electricity consumption influencing factor sequence, and the electricity consumption data of each industrial and commercial user at different times constitutes an electricity data sequence;

[0016] Obtain the sequence to be analyzed, which includes a sequence of electricity consumption influencing factors and a sequence of electricity consumption data;

[0017] Each sequence to be analyzed is divided into several sub-segments; the value at the rightmost end of a sub-segment minus the value at the leftmost end is recorded as the change in the sub-segment; the mean of the changes in all sub-segments of the sequence to be analyzed is recorded as the short-period momentum factor of the sequence to be analyzed, wherein the short-period momentum factor of the sequence to be analyzed includes the short-period momentum factor of the electricity consumption influencing factors and the short-period momentum factor of the electricity consumption data.

[0018] Preferably, the specific steps for ranking the electricity-influencing factors based on the difference in short-period momentum factors between each electricity-influencing factor and the electricity consumption data to obtain a source list of electricity-influencing factors for each industrial and commercial user are as follows:

[0019] Let x0 be the short-period momentum factor of the i-th electricity consumption influencing factor for each industrial and commercial user. i Let x0 be the short-period momentum factor of each industrial and commercial user's electricity consumption data, and let ||x0| be the short-period momentum factor of each user's electricity consumption data. i |-|x0|| is denoted as the source factor of the i-th electricity consumption influencing factor on the electricity consumption data. According to the source factor in descending order, all electricity consumption influencing factors of each industrial and commercial user are sorted to obtain the source list of electricity consumption influencing factors of each industrial and commercial user. || represents taking the absolute value.

[0020] Preferably, the specific steps for tracing the source influence factors of each electricity consumption influencing factor on adjacent electricity consumption influencing factors are as follows:

[0021] In the list of factors affecting electricity consumption, the j-th factor affecting electricity consumption is denoted as y. j The next adjacent electricity consumption influencing factor of the j-th electricity consumption influencing factor is y. j+1 ;y j and y j+1 The sequences formed by the values ​​at different times are the electricity consumption influencing factor sequences Y. j and Y j+1 , for Y j and Y j+1 Y1 is obtained by filtering separately. j and Y1 j+1 Get Y1 j and Y1 j+1 The time difference a1 between the extreme points with the largest values j Then obtain Y1 j and Y1 j+1 The time difference a2 between the second largest extreme points j , will [a1 j a2 j [This is] the source influence factor of the j-th electricity consumption influencing factor on the next adjacent electricity consumption influencing factor.

[0022] Preferably, the electricity consumption influencing factors in the electricity consumption traceability factor list of all industrial and commercial users are merged based on the traceability influencing factors to obtain a merged traceability table and a merged user for each industrial and commercial user. The specific steps include the following:

[0023] Any user is designated as the first user. Any pair of adjacent electricity consumption influencing factors in the electricity consumption traceability factor list of the first user are designated as A and B, where B is the next adjacent electricity consumption influencing factor of A.

[0024] Users other than the first user are referred to as reference users. In the electricity consumption traceability factor list of each reference user, it is determined whether there is a pair of electricity consumption influencing factors that are the same as and adjacent to A and B. If so, the pair of electricity consumption influencing factors that are the same as and adjacent to A and B are referred to as C and D, where D is the next adjacent electricity consumption influencing factor of C.

[0025] Calculate the cosine similarity between the source influence factor γ(A,B) of electricity consumption factor A on the next adjacent electricity consumption factor B and the source influence factor γ(C,D) of electricity consumption factor C on the next adjacent electricity consumption factor D.

[0026] When the cosine similarity is greater than a preset second threshold, the reference user is recorded as the fusion user of the first user, and C and D are recorded as the fusion factors of the fusion user. C and D are inserted before A and B to obtain the fusion traceability table.

[0027] Preferably, the traceability scale is determined by the average insertion interval of the fusion factors in the fusion traceability table.

[0028] Preferably, the specific steps for constructing the wide data table based on the tracing scale are as follows:

[0029] Each industrial and commercial user and all of their integrated users are recorded sequentially in a wide data table; the recorded content includes the electricity consumption data of industrial and commercial users and integrated users, and the electricity consumption influencing factors in the electricity consumption traceability factor list;

[0030] Select F integrated users from all integrated users of each industrial and commercial user, and mark each industrial and commercial user and the F integrated users with the same label in the data wide table, where F represents the traceability scale.

[0031] Preferably, the traced scale is an integer that is negatively correlated with the average insertion interval.

[0032] Preferably, after constructing the wide data table, the method further includes: when any industrial or commercial user's electricity consumption data is abnormal, searching the wide data table for that industrial or commercial user and all converged users with the same tag, and allocating electricity to that industrial or commercial user and all its converged users.

[0033] The beneficial effects of the technical solution of the present invention are:

[0034] This invention obtains a traceability list of electricity influencing factors for each industrial and commercial user based on each electricity influencing factor and the short-period momentum factor of electricity data, ensuring that the electricity consumption of each industrial and commercial user is traceable, which helps to detect the electricity consumption of industrial and commercial users and solve the power distribution problem from the root.

[0035] Further, based on the source list of electricity consumption influencing factors, the source influence factor of each electricity consumption influencing factor on the next electricity consumption influencing factor is obtained. This source influence factor describes the specific impacts of one electricity consumption influencing factor on the next electricity consumption influencing factor.

[0036] Therefore, by integrating different industrial and commercial users based on the source-tracing influencing factors, a fusion source-tracing table can be obtained. This can effectively combine electricity consumption influencing factors that have similar or identical effects, so that both electricity consumption monitoring and power distribution can be analyzed from the overall industrial and commercial layout. This avoids the problem of low power distribution efficiency caused by monitoring and distributing power to only one industrial and commercial user while ignoring other industrial and commercial users.

[0037] Finally, a wide data table is constructed based on the integrated traceability table. This ensures that, in addition to being accurate and traceable, the wide data table also provides an important data foundation for efficient, rapid, and targeted power allocation. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art 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.

[0039] Figure 1 This is a flowchart illustrating the steps of a method for analyzing the electricity consumption characteristics of industrial and commercial users based on a momentum characteristic model, as provided in an embodiment of the present invention. Detailed Implementation

[0040] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the method for analyzing the electricity consumption characteristics of industrial and commercial users based on a momentum characteristic model proposed by the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0041] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0042] The following description, in conjunction with the accompanying drawings, details the specific scheme of the method for analyzing the electricity consumption characteristics of industrial and commercial users based on the momentum characteristic model provided by this invention.

[0043] Example 1:

[0044] Please see Figure 1 The diagram illustrates a flowchart of the steps in the method for analyzing the electricity consumption characteristics of industrial and commercial users based on a momentum characteristic model, as provided in Embodiment 1 of the present invention. The method includes the following steps:

[0045] Step S1: Collect electricity consumption data from industrial and commercial users and obtain all electricity consumption influencing factors for each user.

[0046] The electricity consumption data of each industrial and commercial user is collected using smart meters. In this embodiment, the electricity consumption data refers to the average electricity consumption per week. The electricity consumption data of each industrial and commercial user for all times (i.e. all weeks) in the most recent year constitutes the electricity consumption data sequence.

[0047] Each industrial and commercial user has different attribute information, such as the user's business status, the operation status of power distribution nodes, and the operating time of production equipment (such as equipment for manufacturing products, air conditioning equipment, etc.). These attribute information will more or less affect the electricity consumption data of industrial and commercial users.

[0048] In order to further extract the main factors affecting the electricity consumption data of industrial and commercial users, so as to record and monitor electricity consumption based on these factors, we first quantify the various attribute information of each industrial and commercial user, and obtain the attribute value of each attribute information of each industrial and commercial user. The attribute value of each attribute information is recorded once a week. The attribute values ​​of each attribute information at all times (i.e. all weeks) in the most recent year constitute the attribute sequence of each attribute information.

[0049] This embodiment standardizes the attribute sequence and the electricity consumption data sequence. This embodiment uses the z-score algorithm for standardization. Other embodiments may use other data standardization methods depending on the specific data. This embodiment does not impose any specific limitations.

[0050] Then, based on the electricity consumption data and attribute information of industrial and commercial users, all electricity consumption influencing factors for each industrial and commercial user are obtained. These factors are attribute information that significantly impacts the electricity consumption of industrial and commercial users.

[0051] As an example, the method for obtaining factors affecting electricity consumption is as follows:

[0052] The grey relational analysis method is used to obtain the weight of each attribute based on the attribute sequence of all attribute information of each industrial and commercial user and the electricity consumption data sequence of each industrial and commercial user. The attributes with weights greater than the first threshold th1 are recorded as the electricity consumption influencing factors of each user.

[0053] Grey relational analysis is an existing weighting analysis method used to evaluate the weight scores of attributes. The more significant the impact of attribute information on the electricity consumption of industrial and commercial users, the higher the weight score of the attribute information; conversely, the less significant the impact of attribute information on the electricity consumption of industrial and commercial users, the lower the weight score of the attribute information. Therefore, in this embodiment, attributes with larger weights are used as factors influencing electricity consumption.

[0054] Other weighting methods, such as the analytic hierarchy process, can be used in other embodiments.

[0055] Furthermore, this embodiment uses th1 = 0.45 as an example for description. Other values ​​can be set in other embodiments, and this embodiment does not impose specific limitations.

[0056] Step S2: Sort the electricity influencing factors based on the difference in short-period momentum factors between each electricity influencing factor and the electricity data, and obtain a source list of electricity influencing factors for each industrial and commercial user.

[0057] Step S201: Obtain the short-period momentum factor of each electricity consumption influencing factor and the short-period momentum factor of the electricity consumption data.

[0058] The momentum factor is a feature quantity in the momentum characteristic model used to describe the changing trend of data over a period of time, such as an upward or downward trend. In this embodiment, the short-term momentum factor (abbreviated as short-period momentum factor) is used to describe the changes in electricity consumption influencing factors and electricity consumption data in a short period of time.

[0059] As an example, the method for obtaining the short-period momentum factor is as follows:

[0060] As can be seen from step S1 above, the factors affecting electricity consumption are essentially attribute information. The values ​​of the factors affecting electricity consumption are the attribute values ​​of the attribute information. The values ​​of each factor affecting electricity consumption at different times within a year constitute the sequence of factors affecting electricity consumption.

[0061] Each electricity consumption influencing factor sequence is equally divided into L1 sub-segments, denoted as the first sub-segment; the value at the rightmost end of the first sub-segment of each electricity consumption influencing factor minus the value at the leftmost end is denoted as the change in the first sub-segment; the mean of the changes in all first sub-segments of each electricity consumption influencing factor is denoted as the short-period momentum factor of each electricity consumption influencing factor.

[0062] The electricity consumption data sequence of each industrial and commercial user is divided into L1 segments, which are denoted as the second sub-segment. The change of the second sub-segment is denoted as the value at the rightmost end minus the value at the leftmost end of the second sub-segment for each industrial and commercial user. The mean of the changes of all the second sub-segments for each user is denoted as the short-period momentum factor of the electricity consumption data.

[0063] This embodiment uses L1=8 as an example. When the data is not divided equally, the last segment is considered as a sub-segment regardless of how much data remains.

[0064] Although the division method described in this embodiment is simple, it can be completed quickly when there are a large number of industrial and commercial users and numerous factors affecting electricity consumption. Furthermore, other division methods can be used in other embodiments of the present invention, for example:

[0065] The electricity consumption influencing factor sequence or electricity consumption data sequence is decomposed using the STL algorithm to obtain periodic terms. The sequence is then divided according to the length of each period within the periodic term; that is, the length of each sub-segment is equal to the length of a period. Compared to the division method in this embodiment, this method is more accurate and automated, but it involves a large computational load and is not suitable for situations with a large number of industrial and commercial users and numerous electricity consumption influencing factors. Therefore, this embodiment will not elaborate on the specific process. The STL algorithm, or time series decomposition algorithm, is a well-known technology.

[0066] Step S202: Obtain a source list of electricity consumption influencing factors for each user by sorting the electricity consumption influencing factors according to the short-period momentum factor.

[0067] The electricity consumption of industrial and commercial users is directly or indirectly affected by factors that influence electricity consumption. These factors have a traceable relationship. For example, the operating conditions or average working hours of industrial and commercial enterprises directly affect the operation of production equipment (e.g., operating time, number of failures, number of start-ups and shutdowns). In turn, the operation of production equipment directly affects the electricity consumption of industrial and commercial users. For such traceable factors influencing electricity consumption, their traceability characteristics should be analyzed. Only in this way can the root causes affecting the electricity consumption of industrial and commercial users be monitored, which will help to allocate power according to the root problems and causes during power dispatching.

[0068] Since the short-period momentum factor obtained above describes the changes in electricity consumption influencing factors and electricity consumption data in a short period of time, the difference between the changes in different electricity consumption influencing factors and the changes in electricity consumption data can reflect the tracing relationship between different electricity consumption influencing factors; therefore, this embodiment obtains a tracing list of electricity consumption influencing factors for each user by sorting the electricity consumption influencing factors using the short-period momentum factor. This tracing list of electricity consumption influencing factors is used to describe the tracing order among all electricity consumption influencing factors for each industrial and commercial user, and tracing can be performed according to the order of electricity consumption influencing factors in the tracing list of electricity consumption influencing factors.

[0069] As an example, the method for obtaining the list of factors affecting electricity consumption is as follows:

[0070] Let x0 be the short-period momentum factor of the i-th electricity consumption influencing factor for each user. i Let x0 be the short-period momentum factor of each user's electricity consumption data, and let ||x0| be the short-period momentum factor of each user's electricity consumption data.i |-|x0|| is denoted as the source factor of the i-th electricity consumption influencing factor on the electricity consumption data. According to the source factor in descending order, all electricity consumption influencing factors of each user are sorted to obtain the source list of electricity consumption influencing factors of each user. || represents taking the absolute value.

[0071] The larger the traceability factor of the i-th electricity consumption influencing factor to the electricity consumption data, the greater the difference between the change of the i-th electricity consumption influencing factor and the change of the electricity consumption data, indicating that the i-th electricity consumption influencing factor and the electricity consumption situation have less direct traceability relationship. Conversely, the smaller the traceability factor, the more direct the relationship. Therefore, any pair of adjacent electricity consumption influencing factors A and B in the traceability list of electricity influencing factors indicates that A directly affects B, and B directly or indirectly affects the electricity consumption situation, where B is the next electricity consumption influencing factor of A.

[0072] Step S3: In the source tracing list of electricity consumption influencing factors for each industrial and commercial user, obtain the source tracing influence factor of each electricity consumption influencing factor on adjacent electricity consumption influencing factors based on the difference in the time corresponding to the maximum value of adjacent electricity consumption influencing factors.

[0073] The aforementioned electricity consumption influencing factor tracing list describes the tracing order among all electricity consumption influencing factors for each industrial and commercial user. This embodiment further analyzes the tracing influence characteristics of adjacent electricity consumption influencing factors based on the electricity consumption influencing factor tracing list. This embodiment uses tracing influence factors to describe the specific impacts of one electricity consumption influencing factor on the next, so that the influencing factors that produce the same impact can be merged in the future, so as to achieve the purpose of merging all industrial and commercial user data together to monitor electricity consumption and realize the purpose of all industrial and commercial users as a whole for electricity allocation.

[0074] As an example, the method for obtaining the source impact factor is as follows:

[0075] In the list of factors affecting electricity consumption, the j-th factor affecting electricity consumption is denoted as y. j The next adjacent electricity consumption influencing factor of the j-th electricity consumption influencing factor is y. j+1 ;y j and y j+1 The sequences formed by the values ​​at different times are the electricity consumption influencing factor sequences Y. j and Y j+1 , for Y j and Y j+1 Y1 is obtained by performing Gaussian filtering on each side. j and Y1 j+1 Get Y1 j and Y1 j+1 The time difference a1 between the extreme points with the largest values j Then obtain Y1j and Y1 j+1 The time difference a2 between the second largest extreme points j , will [a1 j a2 j [This is] the source influence factor of the j-th electricity consumption influencing factor on the next adjacent electricity consumption influencing factor.

[0076] a1 j a2 j This describes the lag pattern or the misalignment pattern in the time-varying trend of the series of electricity-influencing factors. Therefore, tracing the source of influencing factors [a1] j a2 j This describes the specific impact of one electricity consumption factor on the next adjacent electricity consumption factor over time.

[0077] In other embodiments, when obtaining the source tracing influence factor, the time difference a3 between the third largest extreme points can also be found using the method described in this embodiment. j The time difference a4 between the fourth and third largest extreme points j ; Construct the source-tracing impact factor [a1] j a2 j a3 j a4 j The present invention will not elaborate on the implementation methods of other embodiments that are the same as the above-described implementation concept;

[0078] Step S4: In the list of electricity consumption traceability factors for all industrial and commercial users, the electricity consumption influencing factors in the list of electricity consumption traceability factors for all users are merged based on the traceability influencing factors to obtain the merged traceability table and the merged user for each user.

[0079] Different industrial and commercial users should be considered as a whole when using and distributing electricity. For example, multiple industrial and commercial users may use the same power supply node, or they may be under the same transformer or the same substation. Furthermore, multiple industrial and commercial users may have the same electricity policies, such as power rationing policies during peak hours. These industrial and commercial users should be analyzed and considered together in terms of power monitoring and distribution. Therefore, this embodiment integrates traceability data of these industrial and commercial users to achieve the purpose of overall power distribution.

[0080] The fusion method is based on the aforementioned source-tracing impact factors. These factors describe the specific impacts of one electricity consumption impact factor on another. When using source-tracing impact factors to fuse different commercial users, it effectively combines electricity consumption impact factors that have similar or identical effects. This allows for analysis of both electricity consumption monitoring and power distribution from the perspective of the overall industrial and commercial layout, avoiding the problem of low power distribution efficiency caused by monitoring and distributing power only to one industrial and commercial user while ignoring other users.

[0081] The fusion traceability table represents the result after merging the factors affecting electricity consumption.

[0082] As an example, the method for obtaining the fusion traceability table is as follows.

[0083] Any user is designated as the first user. Any pair of adjacent electricity consumption influencing factors in the electricity consumption traceability factor list of the first user are designated as A and B, where B is the next adjacent electricity consumption influencing factor of A.

[0084] In other embodiments, the user with the highest average electricity consumption can be designated as the first user, or the user with the most power outages can be designated as the first user.

[0085] Other users besides the first user are referred to as reference users. In turn, in the electricity consumption traceability factor list of each reference user, it is determined whether there is a pair of electricity consumption influencing factors that are the same as and adjacent to A and B. If there is, the same pair of electricity consumption influencing factors are referred to as C and D, where D is the next adjacent electricity consumption influencing factor of C. If not, the search continues to look for other reference users' electricity consumption traceability factor lists.

[0086] Calculate the cosine similarity between the source influence factor γ(A,B) of electricity consumption factor A on the next adjacent electricity consumption factor B and the source influence factor γ(C,D) of electricity consumption factor C on the next adjacent electricity consumption factor D.

[0087] When the cosine similarity is greater than the second threshold th2, the reference user is recorded as the first user's fused user, and C and D are recorded as the fusion factors of the fused user. C and D are inserted before A and B to obtain the fusion source table.

[0088] For example, if one of the adjacent electricity consumption influencing factors of the first user is (A1, B1), and the electricity consumption tracing factor list of the first reference user also includes (A1, B1), in order to distinguish (A1, B1) of the first user from (A1, B1) of the first reference user, this embodiment will use a different symbol representation method for (A1, B1) of the first reference user, and denote them as (C1, D1) respectively.

[0089] When the cosine similarity of (A1, B1) and (C1, D1) is greater than th2, then the first reference user is the first user's fusion user, and the adjacent electricity consumption influencing factors (C1, D1) are fusion factors, and (C1, D1) is inserted before (A1, B1) (in other embodiments, it can be inserted after).

[0090] Next, another adjacent electricity consumption influencing factor of the first user is (A2, B2), and the electricity consumption tracing factor list of the second reference user also includes (A2, B2). In order to distinguish (A2, B2) of the first user from (A2, B2) of the second reference user, this embodiment uses a different symbol representation method for (A2, B2) of the first reference user, and records them as (C2, D2) respectively.

[0091] When the cosine similarity of (A2, B2) and (C2, D2) is greater than th2, then the second reference user is also the fusion user of the first user. At the same time, the adjacent electricity consumption influencing factors (C2, D2) are fusion factors, and (C2, D2) is inserted before (A2, B2) (in other embodiments, it can be inserted after).

[0092] Therefore, there are multiple fusion users in the fusion traceability table, and each fusion user has at least one fusion factor. Thus, the fusion traceability table contains multiple fusion factors, and each fusion factor is inserted at a different position in the fusion traceability table.

[0093] This embodiment uses th2 = 0.3 as an example for description. Other values ​​can be set in other embodiments, and this embodiment does not impose specific limitations.

[0094] For other industrial and commercial users besides the first user and the first user's merged users, repeat the above method in this step: among the other industrial and commercial users, any industrial and commercial user is re-recorded as the first user, and then the first user's merged users are re-obtained, and then the merged traceability table is obtained again; the specific method is the same as above, and will not be described in detail in this embodiment.

[0095] When there is an industrial or commercial user without a merged user, the list of factors affecting their electricity consumption is used as the merged user traceability table.

[0096] S5. Determine the traceability scale based on the position of the fusion factors included in the fusion user in the fusion traceability table, and construct a wide data table based on the traceability scale.

[0097] A wide data table is a data format used to monitor industrial and commercial electricity consumption. The conventional method for constructing a wide data table involves entering the electricity consumption data and various attribute information of industrial and commercial users one by one. However, this embodiment merges electricity consumption influencing factors that have similar or identical effects into a fusion traceability table. Therefore, to improve the efficiency of industrial and commercial electricity consumption monitoring and the rationality of power allocation, the merged and integrated industrial and commercial user data needs to be reflected in the wide data table.

[0098] The specific method is as follows:

[0099] First, determine the tracing scale based on the position of the fusion factors included in the fusion user in the fusion tracing table.

[0100] The traceability scale is used to describe the continuity of the insertion position of the fusion factor in the fusion traceability table. If multiple fusion elements are inserted consecutively in the fusion traceability table, it indicates that there is a very close power consumption influence relationship between the fusion users corresponding to these fusion elements.

[0101] For example, multiple industrial and commercial users under the same power supply node whose electricity consumption is directly affected by equipment operating time have a large traceability scale. In this case, when allocating electricity to one industrial and commercial user, more consideration needs to be given to the other industrial and commercial users. On the other hand, multiple industrial and commercial users under different power supply nodes whose various electricity consumption influencing factors are different have a smaller traceability scale. In this case, when monitoring or allocating electricity to one industrial and commercial user, it is not necessary to consider the other industrial and commercial users.

[0102] As an example, the method for obtaining the traceability scale is as follows:

[0103] In the fusion traceability table, the positions of the fusion factors inserted in all fusion users are marked, and the traceability scale is obtained based on the average insertion interval of these positions. The traceability scale is negatively correlated with the average insertion interval.

[0104] For example, in the fusion source table [C1, D1, A1, B1, C2, D2, A2, B2, C3, D3, ...], (C1, D1), (C2, D2), and (C3, D3) are fusion factors. Their insertion positions are 1, 5, and 9, respectively, since C1's position in the fusion source table is 1, C2's is 5, and C3's is 9. Therefore, the intervals between insertion positions are 5-1 = 4 and 9-5 = 4, respectively. Thus, the average insertion interval is (4+4)÷2 = 4.

[0105] When a fusion factor is inserted into the fusion traceability table, the average insertion interval is equal to the length of the fusion traceability table minus 2; when no fusion factor is inserted into the fusion traceability table, the average insertion interval is equal to the length of the fusion traceability table.

[0106] The method for constructing a wide data table based on the tracing scale is as follows:

[0107] Construct an empty wide data table. For any fusion traceability table, record all industrial and commercial users involved in the fusion traceability table, i.e., the first user and the fusion user of the first user in step S4 above, row by row in each row of the wide data table. The recorded content includes the user's average electricity consumption data for the most recent week and the values ​​of electricity influencing factors in the electricity influencing factor traceability list. In other embodiments, more data can be recorded, such as other attribute information besides electricity influencing factors. This embodiment does not make specific limitations.

[0108] Select F converged users of the first user and mark them; the purpose of marking is that when marked industrial and commercial users are allocated electricity, other industrial and commercial users with the same mark should also be considered for electricity allocation.

[0109] At this point, a wide data table has been constructed based on the fusion traceability table.

[0110] This completes the embodiment.

[0111] Example 2:

[0112] This embodiment provides a method for quantifying various attribute information of each industrial and commercial user:

[0113] In this embodiment, the attribute information of industrial and commercial users includes: the operating status of industrial and commercial users, the maximum weekly power supply of power supply nodes, the total weekly operating time of equipment, the number of weekly start-ups and shutdowns of equipment, the number of weekly equipment failures, the number of employees working at the enterprise each week, the number of customers received by the enterprise each week, the average daily working hours of the enterprise within a week, and power rationing policies. Other attributes may be added in other embodiments, but this embodiment does not impose specific limitations.

[0114] The operating status of industrial and commercial users can be quantified by their total weekly turnover; that is, the attribute value of the operating status of industrial and commercial users is equal to their total weekly turnover.

[0115] In this embodiment, the power rationing policy can be quantified using the maximum daily power rationing amount per week.

[0116] In other embodiments, some attribute information is text data, such as policy restrictions, business conditions, etc. For text data, the attribute information can be quantized into word vectors using the word2vec method, and then PCA can be used to reduce the dimensionality of the word vectors of the attribute information to one dimension, thereby obtaining the attribute value of the attribute information.

[0117] It should be noted that when some attribute information cannot be collected, the attribute information will no longer participate in the calculation of all embodiments of the present invention. When the attribute value of the attribute information is missing, the missing attribute value is interpolated using the linear interpolation method. When the missing attribute value exceeds 40%, the attribute information is considered to be uncollectible.

[0118] This completes the quantification of various attribute information for each industrial and commercial user.

[0119] This embodiment provides a formula for calculating the traceability scale using the average insertion interval at a location:

[0120]

[0121] Where Q is a hyperparameter, and this embodiment uses Q=100 as an example. Other embodiments can be set to other values, and this embodiment does not specifically limit them. D represents the average insertion interval, and F represents the tracing scale. This indicates that evidence is being collected from the next level.

[0122] In this embodiment, F merged users of the first user are selected for labeling, including:

[0123] For all the integrated users of the first user, obtain the average weekly electricity consumption of the integrated users, and select the F integrated users with the highest average electricity consumption (including the first user) and mark them. The reason for selecting the F integrated users with the highest average electricity consumption is that these F integrated users have a large impact on the power grid due to their large electricity consumption. Therefore, it is more necessary to select these F integrated users to share power together to ensure the stability of the power grid.

[0124] In other embodiments, the number of weekly power outages for converged users can be counted, and the F converged users with the highest number of power outages can be selected and marked.

[0125] The labeling method in this embodiment is as follows: assign the same integer label to the selected F fusion users, and use the integer label as a column of the data wide table.

[0126] It should be noted that, as can be seen from step S4 of Embodiment 1 of the present invention, there are multiple first users in the end. In this embodiment, the integer labels marked for the fused users of different first users cannot be the same.

[0127] In other embodiments, color assignment can be used for marking, with all merged users (including the first user) of each first user marked with the same color, and all merged users of different first users marked with different colors.

[0128] After obtaining the wide data table, this embodiment also includes:

[0129] The system monitors the electricity consumption of each industrial and commercial user in the real-time data wide table. For example, it monitors the electricity consumption of the most recent week. When the electricity consumption of an industrial or commercial user in the most recent week increases or decreases by more than 50% compared to the previous week, it indicates that the industrial or commercial user has an abnormal power consumption. Then, it retrieves the list of factors affecting the power consumption of the industrial or commercial user from the data wide table. It finds the last factor affecting the power consumption in the list, which is most likely to be the factor that directly caused the abnormal power consumption. The industrial or commercial user is notified of this factor so that they can make adjustments based on it. For example, if the factor is the number of times the user starts and stops each week, the number of times the user starts and stops each week can be reduced. The specific adjustment method needs to be combined with the specific circumstances of the industrial or commercial user, which will not be elaborated in this embodiment.

[0130] In other embodiments, more traceable factors affecting electricity consumption can be found in the list of factors affecting electricity consumption. For example, the last two factors in the list are, in order, the power restriction policy and the number of times the equipment is started and stopped each week. It is possible that the restriction policy is too strict, resulting in a high number of times the equipment is started and stopped, which leads to abnormal electricity consumption. In this case, an application can be made to relax the power restriction policy.

[0131] At the same time, other industrial and commercial users with the same identifier as the industrial and commercial user are identified, and electricity is allocated based on the specific electricity consumption of these industrial and commercial users.

[0132] The power distribution method is not the problem that this invention aims to solve, so it will not be described in detail in this embodiment.

[0133] In other embodiments, a wide data table can be used to train a neural network model or a machine learning model for power distribution. The specific implementation method will not be described in this embodiment.

[0134] Conventional methods for constructing wide data tables simply involve entering electricity consumption data and various attribute information of industrial and commercial users one by one into the wide data table. This approach fails to reflect the traceability of data and does not integrate or consider the analysis of related data from different industrial and commercial users. In contrast, the wide data tables constructed in all embodiments of this invention can not only quickly trace the root causes affecting the electricity consumption of industrial and commercial users, thus enabling rapid, accurate, and targeted power allocation, but also integrate data from industrial and commercial users that have similar or identical impacts. This ensures more efficient and accurate power allocation based on the wide data table, avoiding the problem of low power allocation efficiency caused by monitoring and allocating power to only one industrial and commercial user while neglecting other users.

[0135] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for analyzing the electricity consumption characteristics of industrial and commercial users based on a momentum characteristic model, characterized in that, The method includes the following steps: Collect electricity consumption data from industrial and commercial users and obtain all electricity consumption influencing factors for each industrial and commercial user; Based on the difference in short-period momentum factors between each electricity consumption influencing factor and electricity consumption data, the electricity consumption influencing factors are ranked to obtain a source list of electricity consumption influencing factors for each industrial and commercial user. In the source tracing list of electricity consumption influencing factors for each industrial and commercial user, the source tracing influence factor of each electricity consumption influencing factor on adjacent electricity consumption influencing factors is obtained based on the difference in the time corresponding to the maximum value of adjacent electricity consumption influencing factors. In the list of electricity consumption traceability factors for all industrial and commercial users, the electricity consumption influencing factors in the list of electricity consumption traceability factors for all industrial and commercial users are merged based on the traceability influencing factors to obtain a merged traceability table and a merged user for each industrial and commercial user, wherein the merged user contains the merged factors; The tracing scale is determined based on the position of the fusion factors included in the fusion user in the fusion tracing table, and a wide data table is constructed based on the tracing scale. The specific steps for obtaining the short-period momentum factor are as follows: For each electricity consumption influencing factor of each industrial and commercial user, the value of each electricity consumption influencing factor at different times constitutes an electricity consumption influencing factor sequence, and the electricity consumption data of each industrial and commercial user at different times constitutes an electricity data sequence; Obtain the sequence to be analyzed, which includes a sequence of electricity consumption influencing factors and a sequence of electricity consumption data; Each sequence to be analyzed is divided into several sub-segments; the value at the rightmost end of a sub-segment minus the value at the leftmost end is recorded as the change in the sub-segment; the mean of the changes in all sub-segments of the sequence to be analyzed is recorded as the short-period momentum factor of the sequence to be analyzed, wherein the short-period momentum factor of the sequence to be analyzed includes the short-period momentum factor of the electricity consumption influencing factors and the short-period momentum factor of the electricity consumption data.

2. The method for analyzing the electricity consumption characteristics of industrial and commercial users based on the momentum characteristic model according to claim 1, characterized in that, The specific steps for obtaining all electricity consumption influencing factors for each industrial and commercial user are as follows: The attribute values ​​of the same attribute for each industrial and commercial user at different times constitute an attribute sequence, and the electricity consumption data of each industrial and commercial user at different times constitute an electricity consumption data sequence. The grey relational analysis method is used to obtain the weight of each attribute based on all the attribute sequences and electricity consumption data sequences of each industrial and commercial user. Attributes with weights greater than a preset first threshold are recorded as electricity consumption influencing factors for each industrial and commercial user.

3. The method for analyzing the electricity consumption characteristics of industrial and commercial users based on the momentum characteristic model according to claim 1, characterized in that, The specific steps involved in ranking electricity consumption influencing factors based on the difference in short-period momentum factors between each electricity consumption influencing factor and electricity consumption data to obtain a source list of electricity consumption influencing factors for each industrial and commercial user are as follows: Let the short-period momentum factor of the i-th electricity consumption influencing factor for each industrial and commercial user be denoted as . The short-period momentum factor of each industrial and commercial user's electricity consumption data is denoted as... ,Will Let be the source factor for the electricity consumption data, which is the i-th electricity consumption influencing factor. Sort all electricity consumption influencing factors for each industrial and commercial user in descending order of the source factor to obtain a source factor list for each industrial and commercial user's electricity consumption influencing factors. This indicates taking the absolute value.

4. The method for analyzing the electricity consumption characteristics of industrial and commercial users based on the momentum characteristic model according to claim 3, characterized in that, The specific steps for tracing the source influence of each electricity consumption influencing factor on adjacent electricity consumption influencing factors are as follows: In the list of factors affecting electricity consumption, the j-th factor affecting electricity consumption is denoted as... The next adjacent electricity consumption influencing factor of the j-th electricity consumption influencing factor is: ; and The sequences formed by the values ​​at different times are the sequences of electricity consumption influencing factors. and ,right and Filtering was performed separately to obtain and , obtain and Time difference between the extreme points with the largest values , then obtain and Time difference between the second largest extreme points ,Will As the source influence factor of the j-th electricity consumption influencing factor on the next adjacent electricity consumption influencing factor.

5. The method for analyzing the electricity consumption characteristics of industrial and commercial users based on the momentum characteristic model according to claim 3, characterized in that, Based on the aforementioned source tracing factors, the electricity consumption influencing factors in the electricity consumption source tracing factor list of all industrial and commercial users are merged to obtain a merged source tracing table and a merged user for each industrial and commercial user. The specific steps include the following: Any user is designated as the first user. Any pair of adjacent electricity consumption influencing factors in the electricity consumption traceability factor list of the first user are designated as A and B, where B is the next adjacent electricity consumption influencing factor of A. Users other than the first user are referred to as reference users. In the electricity consumption traceability factor list of each reference user, it is determined whether there is a pair of electricity consumption influencing factors that are the same as and adjacent to A and B. If so, the pair of electricity consumption influencing factors that are the same as and adjacent to A and B are referred to as C and D, where D is the next adjacent electricity consumption influencing factor of C. Calculate the source influence factor of electricity consumption factor A on the next adjacent electricity consumption factor B. The source-tracing influence factor of electricity consumption factor C on the next adjacent electricity consumption factor D Cosine similarity; When the cosine similarity is greater than a preset second threshold, the reference user is recorded as the fusion user of the first user, and C and D are recorded as the fusion factors of the fusion user. C and D are inserted before A and B to obtain the fusion traceability table.

6. The method for analyzing the electricity consumption characteristics of industrial and commercial users based on the momentum characteristic model according to claim 5, characterized in that, The traceability scale is determined by the average insertion interval of the fusion factors in the fusion traceability table.

7. The method for analyzing the electricity consumption characteristics of industrial and commercial users based on the momentum characteristic model according to claim 1, characterized in that, The specific steps involved in constructing the wide data table based on the traceability scale are as follows: Each industrial and commercial user and all of their integrated users are recorded sequentially in a wide data table; the recorded content includes the electricity consumption data of industrial and commercial users and integrated users, and the electricity consumption influencing factors in the electricity consumption traceability factor list; Select F integrated users from all integrated users of each industrial and commercial user, and mark each industrial and commercial user and the F integrated users with the same label in the data wide table, where F represents the traceability scale.

8. The method for analyzing the electricity consumption characteristics of industrial and commercial users based on the momentum characteristic model according to claim 6, characterized in that, The tracing metric is an integer that is negatively correlated with the average insertion interval.

9. The method for analyzing the electricity consumption characteristics of industrial and commercial users based on the momentum characteristic model according to claim 7, characterized in that, The process of constructing the wide data table also includes: when any industrial or commercial user's electricity consumption data is abnormal, searching the wide data table for that industrial or commercial user and all converged users with the same tag, and allocating electricity to that industrial or commercial user and all its converged users.

Images