Low-voltage distribution network user-transformer relation identification method and system based on multi-dimensional index fusion

Abstract

The invention discloses a low-voltage power distribution network user-transformer relation identification method and system based on multi-dimensional index fusion, and belongs to the technical field of power system data processing and analysis. Sequentially analyzing the synchronism of the user and the distribution transformer on the power abrupt change event through the power abrupt change event, calculating the linear correlation degree, quantifying the load contribution degree of the user to the distribution transformer, and respectively obtaining the event correlation degree, the event coincidence degree, the power reconstruction degree, the global correlation degree and the energy distribution degree; and performing weighted calculation on the five indexes according to a preset index weight to obtain a matching score of the user and the distribution transformer, and identifying the user-transformer relationship between the user and the distribution transformer through the matching score. Therefore, by implementing the application, the problems that the identification precision of the user change relation is easily interfered by the external environment and the user change relation is difficult to adapt to various scenes due to excessive dependence on single electrical data in the prior art can be solved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of power system data processing and analysis, and particularly relates to a low-voltage distribution network user-transformer relationship identification method and system based on multi-dimensional index fusion. BACKGROUND

[0002] Accurate acquisition of the transformer area topology structure of the low-voltage distribution network is a core prerequisite for accurate calculation of line loss, realization of three-phase imbalance management, and rapid fault positioning. The user-transformer relationship, as the electrical connection mapping relationship between the distribution transformer and the user electric meter, identifies the power distribution path by describing which distribution transformer supplies power to which users, and its correctness directly affects the accuracy of the transformer area topology structure.

[0003] Currently, the active power time series data of all users and distribution transformers (hereinafter referred to as "transformers") in the transformer area are collected, the time series synchronization of the power or voltage of the users and the distribution transformers is quantified, and then the users in the transformer area are clustered and grouped to determine the affiliation of the users and the distribution transformers. However, this scheme relies too much on a single power-related index. When the collected data has much noise or a sharp peak load appears in the transformer area, the power synchronization of the users and the distribution transformers will appear a temporary deviation, and the obtained user-transformer relationship has a large error. Therefore, the anti-interference ability of this user-transformer relationship identification method based on a single index is poor and cannot adapt to various complex scenarios. SUMMARY

[0004] The application provides a low-voltage distribution network user-transformer relationship identification method and system based on multi-dimensional index fusion, which can solve the problem in the prior art that the identification accuracy of the user-transformer relationship is easily affected by external environmental interference due to too much reliance on a single electrical data, and the method cannot be applied to various scenarios.

[0005] The first aspect of the application provides a low-voltage distribution network user-transformer relationship identification method based on multi-dimensional index fusion, which comprises:

[0006] Power mutation event identification is performed through a power difference time series to obtain a first event set of each user and a second event set of each distribution transformer; wherein the power difference time series is obtained by differentiating the collected user active power time series and the transformer active power time series;

[0007] Based on the first event set and the second event set, the synchronization of the users and the distribution transformers in the power mutation event is analyzed to obtain an event correlation degree and an event coincidence degree;

[0008] Based on the power difference time series, the linear correlation degree between the users and the distribution transformers and the load contribution degree of the users to the distribution transformers are calculated to obtain a power reconstruction degree, a global correlation degree, and an energy distribution degree, respectively;

[0009] The event correlation degree, the event coincidence degree, the power reconstruction degree, the global correlation degree and the energy distribution degree are weighted calculated according to preset index weights, and the matching scores of each user and all power distribution transformers are obtained, and the user-transformer relationship between the user and the power distribution transformer is determined through the maximum value of the matching scores.

[0010] The above scheme performs difference calculation on the power sequences of the collected users and power distribution transformers, quantifies the change relationship between the powers, and obtains the power difference time sequence. By identifying the abnormal values in the power difference time sequence, the time of the power mutation event is determined, and the related event set is summarized. Then, based on the event sets of the users and the power distribution transformers, the synchronism of the two when the power mutation event occurs and the correlation degree of each power mutation event are analyzed, the correlation between the users and the power distribution transformers is quantified from multiple dimensions, and the corresponding five-dimensional correlation index is obtained. Compared with using power data as a single correlation index, the multi-dimensional correlation index considers the correlation between events and the energy contribution of the equipment, can be applied to scenarios with measurement noise and load fluctuations, fully covers user groups with different power consumption characteristics, and greatly improves the identification accuracy of the user-transformer relationship.

[0011] In a possible implementation method of the first aspect, power mutation event identification is performed through the power difference time sequence, and first event sets of each user and second event sets of each power distribution transformer are obtained, specifically as follows:

[0012] The output powers of each user and each power distribution transformer in the target area are collected to obtain user active power time sequence and power distribution transformer active power time sequence;

[0013] The user active power time sequence and the power distribution transformer active power time sequence are first-order differentiated to obtain the power difference time sequence of the user and the power distribution transformer;

[0014] The absolute values of the differences in the power difference time sequence are extracted, time points with the absolute values of the differences greater than or equal to a preset detection threshold are taken as power mutation events, and first event sets of each user and second event sets of each power distribution transformer are obtained.

[0015] The above scheme collects the output powers of the user group and the power distribution transformer in a period of time to obtain corresponding power time sequence data. The power change relationship between the user and the power distribution transformer is reflected by performing difference calculation on the power time sequence data, and the influence of the static load component is eliminated. The abnormal values in the difference sequence are taken as the time of the power mutation event, which provides data support for subsequent analysis of the synchronism and similarity of the power mutation of the user and the power distribution transformer.

[0016] In a possible implementation method of the first aspect, based on the first event set and the second event set, the synchronization of the user and the power distribution transformer on the power mutation event is analyzed to obtain an event correlation degree and an event coincidence degree, specifically as follows:

[0017] According to the time window corresponding to the first event, the power difference time sequence is translated, and the event synchronization of the user and the power distribution transformer on the first event is analyzed to obtain an event correlation degree of the user and the power distribution transformer.

[0018] By calculating the intersection of the first event set and the second event set, the time synchronization of the user and the power distribution transformer on the power mutation event is quantified to obtain an event coincidence degree of the user and the power distribution transformer.

[0019] The above scheme calculates the intersection of the event set to determine whether the user and the power distribution transformer have experienced the same power mutation event, thereby analyzing the synchronization in the time dimension.

[0020] In a possible implementation method of the first aspect, according to the time window corresponding to the first event, the power difference time sequence is translated, and the event synchronization of the user and the power distribution transformer on the first event is analyzed to obtain an event correlation degree of the user and the power distribution transformer, specifically as follows:

[0021] A plurality of data points of each first event are collected at a preset time interval to obtain a time window of the first event.

[0022] The power difference time sequence is translated by the time window at a first threshold moment, and the Pearson correlation coefficients under different translation amounts are calculated to obtain an event correlation degree of the power distribution transformer and the first event.

[0023] The mean value of all the correlation degrees is calculated to obtain an event correlation degree of each user and each power distribution transformer.

[0024] The above scheme covers the complete change process before and after the power mutation event through the time window. By calculating the Pearson correlation coefficients under different translation amounts, it is determined whether the user and the power distribution transformer have experienced the power mutation event at the same moment, thereby accurately analyzing the event synchronization of the user and the power distribution transformer.

[0025] In a possible implementation method of the first aspect, based on the power difference time sequence, the linear correlation degree between the user and the power distribution transformer and the load contribution degree of the user to the power distribution transformer are calculated to obtain a power reconstruction degree, a global correlation degree, and an energy distribution degree, specifically as follows:

[0026] quantify the load contribution of each user to the distribution transformer by calculating the energy ratio of each differential value in the power difference sequence, to obtain the energy distribution degree of each user and the distribution transformer; wherein if the energy ratio exceeds a preset energy range, the index weight corresponding to the energy distribution degree is reduced;

[0027] Based on the user active power sequence, the distribution transformer active power sequence and the power difference sequence, the Pearson correlation coefficient between the user and the distribution transformer is calculated to obtain the power reconstruction degree and the global correlation degree.

[0028] The above scheme considers the energy contribution degree, improves the identification of users with less power mutation events, covers more types of user groups, and further improves the accuracy of the identification of the relationship between the user and the distribution transformer. By calculating the power reconstruction degree, the global power change correlation between the user and the distribution transformer is quantified, and the influence of negative correlation on the identification result is avoided. Through the global correlation degree, the similarity of the electrical characteristics of the two can be analyzed, and through the similar electrical characteristics, the user and the distribution transformer with a physical connection relationship can be found more quickly and accurately.

[0029] In a possible implementation method of the first aspect, based on the user active power sequence, the distribution transformer active power sequence and the power difference sequence, the Pearson correlation coefficient between the user and the distribution transformer is calculated to obtain the power reconstruction degree and the global correlation degree, specifically:

[0030] The first Pearson correlation coefficient of the power difference sequence is calculated to quantify the linear correlation degree of the global power change between the user and the distribution transformer, and the power reconstruction degree of the user and the distribution transformer is obtained;

[0031] The second Pearson correlation coefficient of the user active power sequence and the distribution transformer active power sequence is calculated, and the first Pearson correlation coefficient and the second Pearson correlation coefficient are weighted and summed to obtain the global correlation degree of the user and the distribution transformer.

[0032] In a possible implementation method of the first aspect, the event correlation degree, the event coincidence degree, the power reconstruction degree, the global correlation degree and the energy distribution degree are weighted calculated according to the preset index weight, to obtain the matching score of each user and all distribution transformers, specifically:

[0033] According to the dimension of the distribution transformer, the event correlation degree, the event coincidence degree, the power reconstruction degree, the global correlation degree and the energy distribution degree are all normalized to obtain a five-dimensional association index;

[0034] According to the preset event factor, the five-dimensional association index is assigned a corresponding index weight;

[0035] Based on the index weight, the five-dimensional association indexes of each user and each distribution transformer are weighted and calculated, and the matching score is obtained by cumulative summation of the weighted calculation results.

[0036] The above scheme introduces an event factor to assign multiple dimensions of association indexes, realizes adaptive adjustment of index weight in the case of insufficient event samples, and covers more application scenarios.

[0037] In a possible implementation method of the first aspect, according to a preset event factor, corresponding index weights are assigned to the five-dimensional association indexes, specifically:

[0038] The event factor is determined according to a ratio of the total number of power mutation events to a preset minimum event number, wherein the minimum event number is the minimum number of power mutation events of a single user;

[0039] According to the event factor, the index weights of the event association degree, the event coincidence degree and the global correlation degree in the five-dimensional association indexes are set;

[0040] Based on the set index summation constraint, the index weights of the power reconstruction degree and the energy distribution degree in the five-dimensional association indexes are assigned through the set index weights.

[0041] The second aspect of the present application provides a low-voltage distribution network transformer relationship identification system based on multi-dimensional index fusion, which comprises: a power mutation identification module, a synchronicity calculation module, an association degree calculation module and a transformer relationship identification module.

[0042] The power mutation identification module is used to identify power mutation events through power difference time series to obtain a first event set of each user and a second event set of each distribution transformer; wherein the power difference time series is obtained by difference calculation on the collected user active power time series and the distribution transformer active power time series.

[0043] The synchronicity calculation module is used to analyze the synchronicity of users and distribution transformers in power mutation events based on the first event set and the second event set, to obtain an event association degree and an event coincidence degree.

[0044] The association degree calculation module is used to calculate the linear correlation degree between users and distribution transformers and the load contribution degree of users to distribution transformers based on the power difference time series, to obtain a power reconstruction degree, a global correlation degree and an energy distribution degree respectively.

[0045] The household and transformer relationship identification module is configured to perform weighted calculation on the event correlation degree, the event coincidence degree, the power reconstruction degree, the global correlation degree and the energy distribution degree according to preset index weights, to obtain a matching score of each user and all distribution transformers, and to determine the household and transformer relationship between the user and the distribution transformer through a maximum value of the matching score.

[0046] The third aspect of the present application provides a terminal device, the device comprising: a terminal device comprising a processor and a memory, the memory storing a computer program, and the processor implementing the steps of any one of the household and transformer relationship identification methods based on multi-dimensional index fusion of the low-voltage distribution network according to the embodiments of the present application when executing the computer program. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0048] Figure 1 is a specific flowchart of a household and transformer relationship identification method based on multi-dimensional index fusion of the low-voltage distribution network according to an embodiment of the present application;

[0049] Figure 2 is a specific structure diagram of a household and transformer relationship identification system based on multi-dimensional index fusion of the low-voltage distribution network according to an embodiment of the present application;

[0050] Figure 3 is a structure diagram of a terminal device according to an embodiment of the present application. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0052] It should be understood that the step numbers used herein are only for the convenience of description, and are not limited to the execution sequence of the steps.

[0053] First Embodiment

[0054] The users under the same power distribution transformer have high synchronization and similarity in the corresponding voltage curves due to the proximity of physical connection and consistency of power supply path, and the active power fluctuation directly affects the voltage, so the synchronization of the voltage can be indirectly inferred by analyzing the correlation of the active power time series data, thereby judging the user-transformer relationship. However, in the urban residential area of the low-voltage power distribution network, the power data collected in the urban residential area has no obvious power characteristics or may have many mutation values due to the characteristics of scattered user power consumption and obvious morning and evening peak load fluctuations, so that the identified user-transformer relationship is quite different from the actual situation, affecting the subsequent power grid operation and maintenance. Therefore, the existing user-transformer relationship identification method based on a single electrical indicator has insufficient anti-interference ability and low scene applicability.

[0055] As shown in Figure 1 To solve the problem in the prior art that the identification accuracy of the user-transformer relationship is easily affected by external environmental interference due to excessive dependence on single electrical data, and is difficult to be applied to various scenes, a first embodiment of the present application provides a specific flowchart of a low-voltage power distribution network user-transformer relationship identification method based on multi-dimensional index fusion. The low-voltage power distribution network user-transformer relationship identification method based on multi-dimensional index fusion of the present embodiment includes steps S1 to S4, which are described in detail as follows:

[0056] Step S1, power mutation event identification is performed through power difference time series to obtain a first event set of each user and a second event set of each power distribution transformer.

[0057] Because all users under the same power distribution transformer can be regarded as being connected to the same "power bus", the voltage at the outlet of the transformer is the common starting point of the bus, so the power fluctuations of all users on the same power distribution transformer are coupled together, and when the power of a user fluctuates instantaneously, the power of other users on the corresponding power distribution transformer and the bus will also fluctuate similarly.

[0058] Based on the above description, it can be known that the power data of users and power distribution transformers having a physical connection relationship exhibit commonality in the time dimension. Therefore, for all users and power distribution transformers in the target area, in the case where all user-transformer relationships are unknown, the output power of each user and each power distribution transformer in the current period can be collected, and the power fluctuation trend can be analyzed to obtain the user active power time series and the power distribution transformer active power time series, thereby providing data support for effectively capturing the correlation characteristics of users and power distribution transformers.

[0059] Based on the user active power time series and the power distribution transformer active power time series, the power mutation events occurring in the current period of all users and power distribution transformers are identified to obtain the core correlation characteristics of users and power distribution transformers.

[0060] Specifically, first-order difference calculation is performed on the user active power time sequence and the distribution transformer active power time sequence to eliminate the influence of static load components and measurement noise (generally, power data fluctuations caused by meter communication interference) in the data acquisition process, to obtain the power difference time sequence of the user and the distribution transformer. Then, according to the set detection threshold, the absolute value of the power difference time sequence is detected, and the time point at which the absolute value is greater than or equal to the detection threshold is determined as the power mutation event point, so as to filter out small fluctuations, and then obtain the first event set of each user and the second event set of each distribution transformer.

[0061] The value of the detection threshold can be dynamically adjusted according to the set minimum power threshold, so as to realize the detection constraint of the mutation event.

[0062] The calculation formula of the power difference time sequence is:

[0063]

[0064] In the formula, ΔP i (t) is the difference value of the power difference time sequence at time t, P i (t) is the power value of the user / distribution transformer i at time t. The difference value is set to 0 at time t=1 to avoid calculation abnormality caused by no pre-data at the initial time.

[0065] The calculation formula of the detection threshold is:

[0066] Th i = max(EMT, ESF·σ(ΔP i ));

[0067] In the formula, Th i is the detection threshold, EMT is the minimum power threshold, which can be set to 0.22 kW; ESF is the weight coefficient, which is 0.85 in the embodiment of the application; and σ is the standard deviation of the difference value.

[0068] Optionally, in other embodiments, the power mutation event points with a time interval of less than 2 minutes are merged, so as to eliminate event fragmentation caused by measurement noise and obtain accurately positioned power mutation events.

[0069] Step S2, based on the first event set and the second event set, the synchronization of the user and the distribution transformer in the power mutation event is analyzed to obtain the event correlation degree and the event coincidence degree.

[0070] In the embodiment of the present application, for each first event of user i, 7 data points are collected in 15 minutes at preset time intervals to construct a corresponding time window to cover the complete change process before and after the event. Considering the influence of power transmission line delay, the power difference time sequence of the distribution transformer is translated by ±2 time points through the time window, the Pearson correlation coefficients under different translation amounts are calculated, and the maximum correlation coefficient is taken as the correlation degree of the power mutation event. Then take the average of the correlation degrees of all first events to get the event correlation degree of user i and distribution transformer j. Through the above scheme, the event correlation degree of each user and each distribution transformer can be obtained, and the value range of the event correlation degree is [0, 1], and the closer to 1, the stronger the event synchronization of the user and the distribution transformer, and it is likely to have associated power mutation events.

[0071] The Pearson correlation coefficient can be used to measure the linear correlation degree between two variables, and is used to quantify the linear correlation between the user and the distribution transformer in the embodiment of the present application.

[0072] The calculation formula of the event correlation degree is:

[0073]

[0074] In the formula, S event is the event correlation degree of user i and distribution transformer j, E i is the first event set, N e,i is the total number of events, and corr() is the Pearson correlation coefficient.

[0075] As another correlation index provided in the embodiment of the present application, the event coincidence degree is used to represent whether the user and the distribution transformer have power mutation events at the same time. The number of intersection elements of the first event set and the second event set is counted, and then the number of intersection elements is divided by the total number of elements of the first event set to obtain the event coincidence degree of the user and the distribution transformer, which quantifies the time synchronization of the user and the distribution transformer on the power mutation event.

[0076] The calculation formula of the event coincidence degree is:

[0077]

[0078] In the formula, S coinc (i,j) is the event coincidence degree of user i and distribution transformer j, E i is the first event set, E j is the second event set, 10 -9 is an adjustment coefficient to avoid the case that the denominator is 0.

[0079] The value range of the event coincidence degree is [0, 1], and the closer to 1, the higher the event time coincidence degree of the user and the distribution transformer.

[0080] Step S3, based on the power difference sequence, calculating the linear correlation degree between the user and the power distribution transformer and quantifying the load contribution of the user to the power distribution transformer, respectively obtaining the power reconstruction degree, the global correlation degree and the energy distribution degree.

[0081] In the embodiments of the present application, the correlation index is also constructed for the power trend, the energy distribution and the global correlation, so as to realize the accurate analysis of the correlation between the user and the power distribution transformer.

[0082] The energy ratio of each difference value in the power difference sequence is calculated to quantify the load contribution of each user to the power distribution transformer, and the energy distribution degree of each user and the power distribution transformer is obtained, specifically: according to the power difference sequence, the energy ratio of the power difference between the user and the power distribution transformer is calculated, and in order to avoid the calculation abnormality caused by the zero power difference of the power distribution transformer, a minimum value is also added in the denominator. If the energy ratio exceeds the pre-set reasonable range, the index weight corresponding to the energy distribution degree can be reduced in the subsequent weighted calculation.

[0083] The calculation formula of the energy distribution degree is:

[0084]

[0085] In the formula, S energy (i,j) is the energy distribution degree of user i and power distribution transformer j.

[0086] The first Pearson correlation coefficient of the power difference sequence between the user and the power distribution transformer is calculated, and the correlation is squared to quantify the global power change correlation between the two, so as to avoid the influence of negative correlation on the relationship identification result of the power distribution transformer.

[0087] The calculation formula of the power reconstruction degree is:

[0088] S recon (i,j) is the power reconstruction degree of user i and power distribution transformer j. i (1:T),ΔP j (1:T)) 2 ;

[0089] In the formula, S recon (i,j) is the power reconstruction degree of user i and power distribution transformer j, and T is the total number of sampling time.

[0090] The value range of the power reconstruction degree is [0, 1], and the larger the value is, the more consistent the global power change trend between the user and the power distribution transformer is.

[0091] In the embodiments of the present application, the second Pearson correlation coefficient between the active power time sequence of the user and the active power time sequence of the distribution transformer (the time sequence contains static load) is calculated, the first Pearson correlation coefficient and the second Pearson correlation coefficient are each given a weight of 0.5 and summed to obtain the global correlation degree of the user and the distribution transformer, and the electrical characteristics in the dynamic and static states are complemented. In other embodiments, the weight value given can be set to other values.

[0092] The calculation formula of the global correlation degree is:

[0093] S global (i,j) = 0.5 corr(P i (1:T), P j (1:T)) + 0.5 corr(ΔP i (1:T), ΔP j (1:T));

[0094] In the formula, S global (i,j) is the global correlation degree of the user i and the distribution transformer j.

[0095] The global correlation degree has a value range of [0, 1], and the larger the value is, the more similar the overall electrical characteristics of the user and the distribution transformer are.

[0096] Step S4, the event correlation degree, the event coincidence degree, the power reconstruction degree, the global correlation degree and the energy distribution degree are weighted and calculated according to the preset index weight, the matching score of each user and all distribution transformers is obtained, and the user-transformer relationship between the user and the distribution transformer is determined through the maximum value of the matching score.

[0097] By constructing the five dimensions of "event correlation-power reconstruction-energy distribution-event coincidence-global correlation", the embodiments of the present application quantify the correlation between the user and the distribution transformer from the dynamic event synchronicity, the global power trend, the load contribution degree and other multi-dimensional indicators. In order to facilitate subsequent user-transformer relationship identification, the multi-dimensional indicators of each user also need to be normalized to obtain standard five-dimensional correlation indicators.

[0098] For each user, the event correlation degree, the event coincidence degree, the power reconstruction degree, the global correlation degree and the energy distribution degree are normalized to the interval [0, 1] according to the dimension of the distribution transformer.

[0099] For example, taking the event correlation degree as an example, the event correlation degree of the user i and the distribution transformer j is subtracted from the minimum value of the event correlation degree of the user i and all distribution transformers, and then divided by the difference between the maximum value and the minimum value of the event correlation degree of the user i and all distribution transformers, to realize the normalization of the event correlation degree. The other indicators also adopt this normalization scheme.

[0100] The normalized event correlation degree is:

[0101]

[0102] In the formula, S' event (i,j) is the normalized event correlation degree of user i and power distribution transformer j, M is the total number of power distribution transformers, and j' is the traversal of all power distribution transformers.

[0103] By normalizing these indicators, the standard five-dimensional correlation indicators are obtained.

[0104] The application embodiment also introduces an "event factor" to allocate the indicator weights of the five-dimensional correlation indicators.

[0105] Specifically, the total number of power mutation events and the minimum number of power mutation events of a single user are counted. If the total number of power mutation events is greater than or equal to 5, the event factor is 1; if the total number of power mutation events is less than 5, the ratio of the total number to the minimum number is taken as the event factor.

[0106] Based on the determined event factor, the indicator weights of the five-dimensional correlation indicators are allocated, and the sum of all indicator weights is required to be 1. According to the event factor, the indicator weights of the event correlation degree, the event coincidence degree and the global correlation degree in the five-dimensional correlation indicators are set; based on the set sum of all indicator weights being 1, the indicator weights of the power reconstruction degree and the energy distribution degree in the five-dimensional correlation indicators are allocated through the set indicator weights.

[0107] Optionally, in the application embodiment, the event factor is set to 1.0, the indicator weight of the event correlation degree is 0.34, the indicator weight of the power reconstruction degree is 0.26, the indicator weight of the energy distribution degree is 0.10, the indicator weight of the event coincidence degree is 0.20, and the weight of the global correlation degree is 0.10.

[0108] Then, for each user i, the five-dimensional correlation indicators of the user i and each power distribution transformer j are multiplied by the corresponding indicator weights, respectively, and the products are added to obtain the matching score of the user i and the power distribution transformer j. By comparing the matching scores of the user i and all power distribution transformers, the power distribution transformer corresponding to the maximum score is taken as the belonging power distribution transformer of the user i, and the identification of the household transformer relationship is completed.

[0109] Compared with a single index, the five-dimensional correlation index provided in the embodiments of the present application can comprehensively capture the multi-dimensional correlation characteristics of users and distribution transformers, accurately identify the user-transformer relationship in load fluctuation scenarios such as peak load periods, and comprehensively cover user groups with different power consumption characteristics, thereby improving the universality of the method. Moreover, the embodiments of the present application do not require additional deployment of hardware devices, but only rely on the existing 15-minute interval measurement data (power, voltage) of smart meters to achieve identification, thereby avoiding the high hardware cost of the signal injection method; at the same time, without manual on-site investigation, the problem of low efficiency and high cost of manual detection is solved, and the method can be quickly applied in large-scale low-voltage distribution networks.

[0110] The identification result of the user-transformer relationship can be used for line loss calculation, three-phase imbalance management and other daily operation and maintenance tasks, and provides accurate topological data support for fine management of distribution networks, which helps to reduce line loss rate, improve three-phase load distribution, and improve the efficiency and economy of distribution network operation.

[0111] The embodiments of the present application have the following beneficial effects:

[0112] The embodiments of the present application quantize the change relationship between powers by performing difference calculation on the collected power sequences of users and distribution transformers, obtain power difference time sequences. By identifying the abnormal values in the power difference time sequences, the time of the power mutation event is determined, and the related event set is summarized. Then, based on the event sets of users and distribution transformers, the synchronization of the two when the power mutation event occurs and the correlation degree of each power mutation event are analyzed, the correlation between users and distribution transformers is quantized from multiple dimensions, and the corresponding five-dimensional correlation index is obtained. Compared with using power data as a single correlation index, the multi-dimensional correlation index considers the correlation between events and the energy contribution of devices, can be applied to scenarios with measurement noise and load fluctuations, can comprehensively cover user groups with different power consumption characteristics, and can greatly improve the identification accuracy of user-transformer relationship.

[0113] Second embodiment

[0114] Further, in order to execute the low-voltage distribution network user-transformer relationship identification system based on multi-dimensional index fusion corresponding to the above-mentioned method embodiments, to realize the corresponding functions and technical effects, Figure 2 A structural diagram of a low-voltage distribution network user-transformer relationship identification system based on multi-dimensional index fusion is provided. For ease of illustration, only the part related to the present embodiment is shown, and the low-voltage distribution network user-transformer relationship identification system based on multi-dimensional index fusion provided by the embodiments of the present application comprises:

[0115] The power mutation identification module 201 is configured to identify power mutation events by power difference time sequence to obtain a first event set of each user and a second event set of each distribution transformer; wherein the power difference time sequence is obtained by differentiating the collected active power time sequence of the user and the active power time sequence of the distribution transformer.

[0116] In the embodiment of the present application, the output power of each user and each distribution transformer in the target table area is collected to obtain the active power time sequence of the user and the active power time sequence of the distribution transformer.

[0117] The first-order difference of the active power time sequence of the user and the active power time sequence of the distribution transformer is calculated to obtain the power difference time sequence of the user and the distribution transformer.

[0118] The absolute value of the difference in the power difference time sequence is extracted, and the time point at which the absolute value of the difference is greater than or equal to a preset detection threshold is taken as a power mutation event to obtain the first event set of each user and the second event set of each distribution transformer.

[0119] The synchronism calculation module 202 is configured to analyze the synchronism of the user and the distribution transformer in the power mutation event based on the first event set and the second event set to obtain an event correlation degree and an event coincidence degree.

[0120] In the embodiment of the present application, for each first event of the user i, a corresponding time window is constructed by taking 7 data points collected at a preset time interval within 15 minutes to cover the complete change process before and after the event. Considering the influence of the transmission line delay, the power difference time sequence of the distribution transformer is translated by ±2 time points through the time window, the Pearson correlation coefficient under different translation amounts is calculated, and the maximum value of the correlation coefficient is taken as the correlation degree of the power mutation event. Then the average of the correlation degrees of all first events is taken to obtain the event correlation degree of the user i and the distribution transformer j. Through the above scheme, the event correlation degree of each user and each distribution transformer can be obtained, and the value range of the event correlation degree is [0, 1]. The closer to 1, the stronger the event synchronism of the user and the distribution transformer, and it is likely that the associated power mutation event has occurred.

[0121] The Pearson correlation coefficient can be used to measure the linear correlation between two variables, and is used to quantify the linear correlation between the user and the distribution transformer in the embodiment of the present application.

[0122] As another correlation index provided in the embodiment of the present application, the event coincidence degree is used to represent whether the user and the distribution transformer have occurred power mutation events at the same time. The number of intersection elements of the first event set and the second event set is counted, and then the number of intersection elements is divided by the total number of elements of the first event set to obtain the event coincidence degree of the user and the distribution transformer, which quantifies the time synchronism of the user and the distribution transformer in the power mutation event.

[0123] The calculation formula of the event coincidence degree is:

[0124]

[0125] In the formula, S coinc (i,j) is the event coincidence degree of user i and power distribution transformer j, E i is the first event set, E j is the second event set, 10 -9 is an adjustment coefficient, to avoid the case that the denominator is 0.

[0126] The event coincidence degree has a value range of [0, 1], and the closer the value is to 1, the higher the event time coincidence degree of the user and the power distribution transformer.

[0127] The correlation degree calculation module 203 is configured to calculate the linear correlation degree between the user and the power distribution transformer and quantify the load contribution degree of the user to the power distribution transformer based on the power difference time sequence, and obtain a power reconstruction degree, a global correlation degree and an energy distribution degree.

[0128] In the embodiment of the application, the energy ratio of each difference value in the power difference time sequence is calculated to quantify the load contribution degree of each user to the power distribution transformer, and the energy distribution degree of each user and the power distribution transformer is obtained; if the energy ratio exceeds a preset energy range, the index weight corresponding to the energy distribution degree is reduced.

[0129] Based on the user active power time sequence, the power distribution transformer active power time sequence and the power difference time sequence, the Pearson correlation coefficient between the user and the power distribution transformer is calculated to obtain the power reconstruction degree and the global correlation degree.

[0130] The user transformer relationship identification module 204 is configured to perform weighted calculation on the event correlation degree, the event coincidence degree, the power reconstruction degree, the global correlation degree and the energy distribution degree according to a preset index weight, to obtain a matching score of each user and all power distribution transformers, and determine the user transformer relationship between the user and the power distribution transformer through the maximum value of the matching score.

[0131] In some embodiments, the power mutation identification module 201 specifically includes:

[0132] For all users and power distribution transformers in the target area, in the case of unknown all user transformer relationships, the output power of each user and each power distribution transformer in the current period and the power fluctuation trend are analyzed to obtain the user active power time sequence and the power distribution transformer active power time sequence, to provide data support for effectively capturing the association characteristics of the user and the power distribution transformer.

[0133] Based on the user active power time sequence and the distribution transformer active power time sequence, power mutation events of all users and distribution transformers in the current time period are identified, and core association features of the users and the distribution transformers are obtained.

[0134] Specifically, first-order difference calculation is performed on the user active power time sequence and the distribution transformer active power time sequence to eliminate the influence of static load components and measurement noise (generally, power data fluctuations caused by meter communication interference) in the data collection process, and power difference time sequences of the users and the distribution transformers are obtained. Then, the detection threshold is set to detect the difference absolute values in the power difference time sequences, and the time points at which the difference absolute values are greater than or equal to the detection threshold are determined as power mutation event points, so as to filter out small fluctuations, and then first event sets of the users and second event sets of the distribution transformers are obtained.

[0135] The value of the detection threshold can be dynamically adjusted according to the set minimum power threshold, so as to realize the detection constraint of the mutation events.

[0136] The calculation formula of the power difference time sequence is as follows:

[0137]

[0138] In the formula, ΔP i (t) is the difference value of the power difference time sequence at time t, P i (t) is the power value of the user / distribution transformer i at time t. The difference value at time t=1 is set to 0 to avoid calculation abnormalities caused by no pre-data at the initial time.

[0139] The calculation formula of the detection threshold is as follows:

[0140] Th i =max(EMT,ESF·σ(ΔP i ));

[0141] In the formula, Th i is the detection threshold, EMT is the minimum power threshold, which can be set to 0.22 kW; ESF is a weight coefficient, and the value of the weight coefficient in the embodiments of the present application is 0.85; and σ is the standard deviation of the difference value.

[0142] Optionally, in other embodiments, power mutation event points with a time interval of less than 2 minutes are merged, so as to eliminate event fragmentation caused by measurement noise and obtain accurately positioned power mutation events.

[0143] In some embodiments, the association degree calculation module 203 specifically comprises:

[0144] The energy ratio of each differential value in the power difference sequence is calculated to quantify the load contribution of each user to the distribution transformer, and the energy distribution degree of each user and the distribution transformer is obtained. Specifically, the energy ratio of the power difference between the user and the distribution transformer is calculated according to the power difference sequence, and a very small value is added to the denominator to avoid calculation abnormality caused by the power difference of the distribution transformer being zero. If the energy ratio exceeds the preset reasonable range, the index weight corresponding to the energy distribution degree can be reduced in subsequent weighted calculation.

[0145] The calculation formula of the energy distribution degree is:

[0146]

[0147] In the formula, S energy (i,j) is the energy distribution degree of the user i and the distribution transformer j.

[0148] The first Pearson correlation coefficient of the power difference sequence between the user and the distribution transformer is calculated, and the correlation is squared to quantify the global power change correlation between the user and the distribution transformer, so as to avoid the influence of negative correlation on the relationship identification result of the distribution transformer.

[0149] The calculation formula of the power reconstruction degree is:

[0150] S recon (i,j) is the power reconstruction degree of the user i and the distribution transformer j, and T is the total number of sampling time points. i (1:T),ΔP j (1:T)) 2 ;

[0151] In the formula, S recon (i,j) is the power reconstruction degree of the user i and the distribution transformer j, and T is the total number of sampling time points.

[0152] The power reconstruction degree has a value range of [0, 1], and the larger the value is, the more consistent the global power change trend between the user and the distribution transformer is.

[0153] In the embodiment of the application, the second Pearson correlation coefficient between the active power sequence of the user and the active power sequence of the distribution transformer (the sequence contains static load) is calculated, the first Pearson correlation coefficient and the second Pearson correlation coefficient are each given a weight of 0.5 and summed to obtain the global correlation degree of the user and the distribution transformer, so as to realize the complementation of electrical characteristics in dynamic and static. In other embodiments, the weight value can be set to other values.

[0154] The calculation formula of the global correlation degree is:

[0155] S global (i,j) = 0.5·corr(P i (1:T),Pj (1:T))+0.5·corr(ΔP i (1:T),ΔP j (1:T));

[0156] In the formula, S global (i,j) represents the global correlation between user i and distribution transformer j.

[0157] The global correlation value ranges from [0,1], and the larger the value, the more similar the overall electrical characteristics of the user and the distribution transformer are.

[0158] In some embodiments, the household change relationship identification module 204 specifically comprises:

[0159] By constructing five dimensions—"event correlation, power reconfiguration, energy allocation, event overlap, and global correlation"—this application embodiment quantifies the correlation between users and distribution transformers from multiple dimensions, including dynamic event synchronization, global power trends, and load contribution. To facilitate subsequent user-transformer relationship identification, the multi-dimensional indicators for each user also need to be normalized to obtain standard five-dimensional correlation indicators.

[0160] For each user, the event correlation, event overlap, power reconfiguration, global correlation, and energy allocation are normalized to the [0,1] interval according to the dimension of the distribution transformer.

[0161] For example, taking event correlation as an example, the event correlation is normalized by subtracting the minimum event correlation between user i and all distribution transformers from the event correlation between user i and distribution transformer j, and then dividing by the difference between the maximum and minimum event correlation between user i and all distribution transformers. This normalization scheme is used for all other indicators.

[0162] The normalized correlation of the events is:

[0163]

[0164] In the formula, S' event (i,j) represents the normalized event correlation between user i and distribution transformer j, M represents the total number of distribution transformers, and j' represents the number of distribution transformers traversed.

[0165] By normalizing these indicators, standard five-dimensional correlation indicators are obtained.

[0166] This application also introduces an "event factor" to allocate the index weights of the five-dimensional correlation indicators.

[0167] Specifically, the total number of power surge events and the minimum number of power surge events occurring for a single user are counted. If the total number of power surge events is ≥5, the event factor is set to 1; if the total number of power surge events is <5, the ratio of the total number to the minimum number is used as the event factor value.

[0168] Based on the determined event factors, the weights of the five-dimensional correlation indicators are assigned, and the sum of all indicator weights is required to be 1. According to the event factors, the weights of the event correlation degree, event overlap degree, and global relevance among the five-dimensional correlation indicators are set; based on the set sum of all indicator weights being 1, the weights of the power reconfiguration degree and energy allocation degree among the five-dimensional correlation indicators are assigned using the set indicator weights.

[0169] Optionally, in this embodiment of the application, if the event factor is set to 1.0, then the weight of the event correlation index is 0.34, the weight of the power reconfiguration index is 0.26, the weight of the energy allocation index is 0.10, the weight of the event overlap index is 0.20, and the weight of the global correlation index is 0.10.

[0170] Then, for each user i, the five-dimensional correlation index between user i and each distribution transformer j is multiplied by the corresponding index weight, and the multiplications are summed to obtain the matching score between user i and distribution transformer j. By comparing the matching scores between user i and all distribution transformers, the distribution transformer with the highest score is taken as the assigned distribution transformer of user i, thus completing the identification of the user-transformer relationship.

[0171] Compared to a single indicator, the five-dimensional correlation indicator provided in this application embodiment can comprehensively capture the multi-dimensional correlation characteristics between users and distribution transformers, enabling accurate identification of user-transformer relationships even under load fluctuation scenarios such as peak load periods. It can also comprehensively cover user groups with different electricity consumption characteristics, improving the method's versatility. Furthermore, this application embodiment requires no additional hardware deployment; identification can be achieved solely based on the existing 15-minute interval measurement data (power, voltage) from smart meters, avoiding the high hardware costs of signal injection methods. Simultaneously, it eliminates the need for manual on-site inspections, solving the problems of low efficiency and high cost associated with manual detection methods, and can be rapidly promoted and applied in large-scale low-voltage distribution networks.

[0172] The identification results of the household-transformer relationship can be used for routine operation and maintenance tasks such as line loss calculation and three-phase imbalance management, providing accurate topological data support for the refined management of the distribution network, which helps to reduce the line loss rate, improve the three-phase load distribution, and enhance the operating efficiency and economy of the distribution network.

[0173] Implementing the embodiments of this application has the following beneficial effects:

[0174] This application embodiment quantifies the power variation relationship by performing differential calculations on the collected power sequences of users and distribution transformers, obtaining a power differential time series. By identifying outliers in the power differential time series, the timing of power mutation events is determined, and a set of related events is summarized. Then, based on the event sets of users and distribution transformers, the synchronicity of the two at the time of power mutation events and the correlation of each power mutation event are analyzed, quantifying the correlation between users and distribution transformers from multiple dimensions, and obtaining corresponding five-dimensional correlation indicators. Compared with using power data as a single correlation indicator, the multi-dimensional correlation indicator considers the correlation between events and the energy contribution of equipment, is applicable to scenarios with measurement noise and load fluctuations, comprehensively covers user groups with different electricity consumption characteristics, and significantly improves the accuracy of identifying user-transformer relationships.

[0175] Furthermore, Figure 3 This is a structural diagram of a terminal device provided in one embodiment of this application. Figure 3 As shown, the terminal device 3 of this embodiment includes: at least one processor 30 (in... Figure 3 The present invention includes a memory 31 and a computer program 32 stored in the memory 31 and executable on the at least one processor. When the processor 30 executes the computer program 32, it can implement the steps of the low-voltage distribution network household transformer relationship identification method based on multi-dimensional index fusion as described in any one of the embodiments of this application.

[0176] The terminal device 3 may be a computing device such as a desktop computer, a cloud server, or a laptop computer, and the computing device may include, but is not limited to, a processor 30 and a memory 31. Figure 3 This is merely an example of terminal device 3 and does not constitute a limitation on terminal device 3. It may include more or fewer components than those shown in the figure.

[0177] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above descriptions are merely specific embodiments of this application and are not intended to limit the scope of protection of this application. In particular, it should be noted that any modifications, equivalent substitutions, or improvements made by those skilled in the art within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for identifying the relationship between household transformers in a low-voltage distribution network based on multi-dimensional index fusion, characterized in that, include: Power mutation events are identified by power differential timing, resulting in a first event set for each user and a second event set for each distribution transformer; wherein, the power differential timing is obtained by differential calculation of the collected user active power timing and distribution transformer active power timing. Based on the first event set and the second event set, the synchronization between users and distribution transformers on power mutation events is analyzed to obtain the event correlation degree and event overlap degree. Based on the power differential timing, the linear correlation between users and distribution transformers is calculated and the load contribution of users to distribution transformers is quantified, resulting in power reconfiguration degree, global correlation degree, and energy allocation degree, respectively. The event correlation, event overlap, power reconfiguration, global correlation, and energy allocation are weighted according to preset index weights to obtain the matching score between each user and all distribution transformers. The user-transformer relationship between the user and the distribution transformer is determined by the maximum value of the matching score.

2. The method for identifying the relationship between household transformers in a low-voltage distribution network based on multi-dimensional index fusion as described in claim 1, characterized in that, The process of identifying power surge events through power differential timing to obtain a first event set for each user and a second event set for each distribution transformer is as follows: Within the target distribution area, the output power of each user and each distribution transformer is collected to obtain the active power time sequence of the user and the active power time sequence of the distribution transformer. The user's active power time sequence and the distribution transformer's active power time sequence are differentially analyzed to obtain the power differential time sequence between the user and the distribution transformer. Extract the absolute value of the difference in the power differential time series, and take the time point when the absolute value of the difference is greater than or equal to the preset detection threshold as the power mutation event, to obtain the first event set for each user and the second event set for each distribution transformer.

3. The method for identifying the relationship between household transformers in a low-voltage distribution network based on multi-dimensional index fusion as described in claim 1, characterized in that, Based on the first event set and the second event set, the synchronicity between users and distribution transformers during power surge events is analyzed to obtain event correlation and event overlap. Specifically: The power differential timing sequence is shifted according to the time window corresponding to the first event, and the event synchronization between the user and the distribution transformer on the first event is analyzed to obtain the event correlation degree between the user and the distribution transformer. By calculating the intersection of the first event set and the second event set, the time synchronization between the user and the distribution transformer on power surge events is quantified, and the event overlap between the user and the distribution transformer is obtained.

4. The method for identifying the relationship between household transformers in a low-voltage distribution network based on multi-dimensional index fusion as described in claim 3, is characterized in that, The process involves shifting the power differential time sequence according to the time window corresponding to the first event, analyzing the event synchronicity between the user and the distribution transformer at the first event, and obtaining the event correlation degree between the user and the distribution transformer. Specifically: Several data points for each of the first events are collected at preset time intervals to obtain the time window of the first event; The power differential time series is shifted by the time window at a first threshold time, and the Pearson correlation coefficient is calculated under different shift amounts to obtain the correlation degree between the distribution transformer and the first event. Calculate the mean of all the aforementioned correlation degrees to obtain the event correlation degree between each user and each distribution transformer.

5. The method for identifying the relationship between household transformers in a low-voltage distribution network based on multi-dimensional index fusion as described in claim 1, characterized in that, Based on the power differential time series, the linear correlation between users and distribution transformers is calculated, and the load contribution of users to distribution transformers is quantified, yielding the power reconfiguration degree, global correlation degree, and energy allocation degree, respectively. The energy ratio of each differential value in the power differential time series is calculated to quantify the load contribution of each user to the distribution transformer, thereby obtaining the energy allocation degree between each user and the distribution transformer; wherein, if the energy ratio exceeds a preset energy range, the index weight corresponding to the energy allocation degree is reduced. Based on the user's active power time series, the distribution transformer's active power time series, and the power differential time series, the power reconfiguration degree and global correlation degree are obtained by calculating the Pearson correlation coefficient between the user and the distribution transformer.

6. The method for identifying the relationship between household transformers in a low-voltage distribution network based on multi-dimensional index fusion according to claim 5, characterized in that, Based on the user's active power time series, the distribution transformer's active power time series, and the power differential time series, the power reconfiguration degree and global correlation degree are obtained by calculating the Pearson correlation coefficient between the user and the distribution transformer, specifically as follows: The power reconfiguration degree between the user and the distribution transformer is obtained by calculating the first Pearson correlation coefficient of the power differential time series to quantify the linear correlation between the user and the global power change of the distribution transformer. Calculate the second Pearson correlation coefficient between the user's active power time series and the distribution transformer's active power time series, and then perform a weighted sum of the first and second Pearson correlation coefficients to obtain the global correlation between the user and the distribution transformer.

7. The method for identifying the relationship between household transformers in a low-voltage distribution network based on multi-dimensional index fusion according to claim 1, characterized in that, The matching score between each user and all distribution transformers is obtained by weighting the event correlation, event overlap, power reconfiguration, global relevance, and energy allocation according to preset index weights. Specifically: Based on the dimensions of distribution transformers, the event correlation degree, event overlap degree, power reconfiguration degree, global correlation degree, and energy allocation degree are all normalized to obtain five-dimensional correlation indicators; Based on preset event factors, assign corresponding indicator weights to the five-dimensional correlation indicators; Based on the aforementioned index weights, the five-dimensional correlation indexes between each user and each distribution transformer are weighted and calculated, and the matching score is obtained by summing the weighted calculation results.

8. The method for identifying the relationship between household transformers in a low-voltage distribution network based on multi-dimensional index fusion according to claim 7, characterized in that, The step of assigning corresponding index weights to the five-dimensional correlation indicators based on preset event factors is as follows: The event factor is determined based on the ratio of the total number of power surge events to a preset minimum number of events; wherein the minimum number of events is the minimum number of power surge events that occur for a single user. Based on the event factors, set the weights of the event correlation degree, event overlap degree, and global relevance in the five-dimensional correlation indicators; Based on the set total constraint of the indicators, the indicator weights of power reconfiguration degree and energy allocation degree in the five-dimensional correlation indicators are allocated by setting the indicator weights.

9. A low-voltage distribution network customer-transformer relationship identification system based on multi-dimensional index fusion, characterized in that, include: Power mutation identification module, synchronization calculation module, correlation calculation module, and household transformer relationship identification module; The power mutation identification module is used to identify power mutation events through power differential timing to obtain a first event set for each user and a second event set for each distribution transformer. The power differential timing is calculated by differentially calculating the collected user active power timing and the distribution transformer active power timing. The synchronization calculation module is used to analyze the synchronization between users and distribution transformers on power mutation events based on the first event set and the second event set, and to obtain the event correlation degree and event overlap degree. The correlation calculation module is used to calculate the degree of linear correlation between users and distribution transformers and quantify the load contribution of users to distribution transformers based on the power differential time series, so as to obtain the power reconfiguration degree, global correlation degree and energy allocation degree respectively. The user-transformer relationship identification module is used to perform weighted calculations on event correlation, event overlap, power reconfiguration, global correlation, and energy allocation based on preset index weights to obtain the matching score between each user and all distribution transformers. The user-transformer relationship between the user and the distribution transformer is determined by the maximum value of the matching score.

10. A terminal device, characterized in that, It includes a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the steps of the low-voltage distribution network household transformer relationship identification method based on multi-dimensional index fusion as described in any one of claims 1 to 8.

Images