Low-voltage distribution area user phase sequence identification method based on hybrid algorithm framework

Abstract

The invention discloses a low-voltage distribution area user phase sequence identification method based on a hybrid algorithm framework. The method comprises the following steps: 1) calculating an active power sum change sequence of each user cluster; 2) calculating the average absolute variability MAV of each user cluster based on the change sequence Ci; 3) performing descending sort on the user clusters based on the size of the average absolute variability MAV; 4) calculating an active power change sequence of each phase of the transformer based on the user cluster G'g; 5) calculating the correlation r between the active power sum change of the user cluster G'g and the active power change sequence of each phase of the transformer; and if the correlation is maximum and the active power of the phase of the transformer is greater than the sum of the active power of the user cluster G'g, taking the phase as the phase sequence to which the current user cluster G'g belongs. According to the method, dependence on single data quality is overcome through an algorithm fusion framework, data can be analyzed more comprehensively, the recognition accuracy and reliability are improved, and user phase sequence recognition with higher precision is achieved.

Description

Technical Field

[0001] The present invention relates to the field of low-voltage distribution network phase sequence identification, and in particular to a low-voltage distribution network user phase sequence identification method based on a hybrid algorithm framework. Background Art

[0002] Phase sequence identification of low-voltage distribution network is of great significance for improving the operation efficiency, safety and reliability of power grid, as well as supporting the modernization and intelligent management of power grid. At present, in order to obtain a clear user phase sequence relationship, there are mainly methods based on carrier communication, methods based on voltage sequence clustering analysis, methods based on correlation analysis, and methods based on linear integer programming. After research, the inventor found that the prior art has at least the following disadvantages:

[0003] The method based on carrier communication uses the meter to send the corresponding signal, and uses the meter to send the corresponding signal, and uses the received response signal to identify the phase sequence of the user. The advantage of this method is that the user's phase sequence information can be directly obtained, but it has high requirements for the communication and signal processing of the equipment, and may require additional equipment installation and maintenance, and the cost is relatively high. Voltage sequence clustering analysis mainly relies on the similarity of voltage fluctuations to cluster users to identify the phase sequence. It usually uses unsupervised clustering algorithms such as k-means or density-based clustering algorithms (such as DBSCAN). When processing voltage data, these methods need to consider the timing characteristics and fluctuation patterns of the data to improve the accuracy of clustering. However, the voltage fluctuation curve is easily affected by changes in user load, which may affect the stability and accuracy of the clustering results. The phase sequence identification method based on voltage correlation and active power correlation identifies the phase sequence by analyzing the correlation between electrical quantities such as voltage and active power. The advantage of this method is that data is easy to obtain, but it may be affected by data quality, such as missing data or outliers, which may reduce the accuracy of identification. In addition, the linear programming modeling method describes the phase sequence relationship of users by establishing a mathematical model and using an integer programming algorithm to solve it. This method can provide a new perspective to deal with the phase sequence identification problem, but as the number of users to be identified increases, the solution of the model may become complex and computationally intensive, requiring high computing resources. Summary of the invention

[0004] The purpose of the present invention is to provide a method for identifying phase sequence of users in a low-voltage distribution station area based on a hybrid algorithm framework, comprising the following steps:

[0005] 1) Divide the low voltage distribution area user cluster G = {G1, G2, ..., G j};

[0006] 2) Calculate the total active power variation sequence of each user cluster G' = {G1', G'2, ..., G' j};

[0007] 3) Based on the change sequence C i , calculate the average absolute variability MAV of each user cluster;

[0008] 4) Based on the mean absolute variability MAV, the user clusters are sorted in descending order, and the low-voltage distribution station area user cluster is updated, which is recorded as G'={G1',G'2,...,G' j};

[0009] 5) Based on user cluster G' g , calculate the active power change sequence C of each phase of the transformer T,k ;

[0010] 6) Calculate user cluster G' g The correlation r between the total active power change and the active power change sequence of each phase of the transformer; if the correlation is the largest and the active power of the transformer phase is greater than the user cluster G' g The total active power is the current user cluster G' g Phase sequence; the initial value of g is 1;

[0011] 7) Determine whether g>j is established. If not, return to step 5). If so, end the phase sequence identification and output the identification result.

[0012] Furthermore, in step 1), dividing the low-voltage distribution station area into user clusters means dividing the low-voltage distribution station area into user clusters based on the principle of voltage similarity.

[0013] Furthermore, the voltage similarity principle means that the voltage fluctuations of users belonging to the same phase sequence show similarity.

[0014] Further, in step 1), the step of dividing the low-voltage distribution station area user clusters includes:

[0015] 1.1) Cluster the n users in the low-voltage distribution area into j user clusters, denoted by G = {G1, G2, …, G j};

[0016] 1.2) The zero-power user is classified into the user cluster of non-zero-power users whose voltage fluctuation similarity with the zero-power user is greater than a certain value, and the user cluster G is updated.

[0017] Furthermore, in step 1.1), the method of clustering n users in the low-voltage distribution station area into j user clusters includes a PAM clustering algorithm based on the minimum Davis-Boulting index.

[0018] Furthermore, in step 2), the total active power change sequence of each user cluster is recorded as

[0019] The total active power change of each user cluster is as follows:

[0020]

[0021] In the formula, is the total active power of user i at different times.

[0022] Further, in step 3), the mean absolute variability MAV is as follows:

[0023]

[0024] Where M is the length of the active sequence.

[0025] Further, in step 5), the active power change sequence of each phase of the transformer is recorded as

[0026] Among them, the active power change c of the transformer k phase t (P T,k ) is as follows:

[0027]

[0028]

[0029] Where, t = 1, 2, ..., T M ; T M is the number of sample sequences between two different moments; is the active power of the transformer k phase; P Loss,k is the line loss of the transformer k phase; h i Represents user cluster G' g The decision variable of whether user i belongs to phase sequence k; k = {A, B, C} represents the phase sequence; n is the number of users in the current low-voltage distribution station area; m represents the time.

[0030] Further, in step 6), the correlation r is as follows:

[0031]

[0032] In the formula, For user cluster G' g The mean of the total active power change sequence and the active power change sequence of each phase of the transformer.

[0033] The technical effect of the present invention is undoubted, and the beneficial effects of the present invention are as follows:

[0034] 1) The present invention does not require additional equipment installation, which can significantly reduce costs. At the same time, the present invention reduces the need for equipment maintenance and reduces long-term operating costs.

[0035] 2) The present invention does not need to wait for the installation and configuration of new equipment and can be implemented quickly, thereby increasing the startup speed of the project.

[0036] 3) The present invention does not rely on high signal processing and communication functions, and can adapt to environments with poor signals. The present invention can be applied under a variety of network conditions, and can even be applied to areas with imperfect network infrastructure.

[0037] 4) The present invention overcomes the reliance on single data quality through an algorithm fusion framework, can analyze data more comprehensively, improve the accuracy and reliability of recognition, and achieve higher-precision user phase sequence identification.

[0038] 5) The present invention can be integrated into the existing power monitoring system without large-scale system reconstruction. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 Schematic diagram of voltage fluctuation between different phase sequences in a low-voltage area

[0040] Figure 2 This is a schematic diagram of voltage fluctuation curves for some users in a certain phase;

[0041] Figure 3 It is a flow chart of the phase sequence identification method of the present invention. DETAILED DESCRIPTION

[0042] The present invention is further described below in conjunction with the embodiments, but it should not be understood that the above subject matter of the present invention is limited to the following embodiments. Without departing from the above technical ideas of the present invention, various substitutions and changes are made according to the common technical knowledge and customary means in the art, which should all be included in the protection scope of the present invention.

[0043] Embodiment 1:

[0044] See also Figures 1 to 3 , a method for identifying the phase sequence of users in a low-voltage distribution area based on a hybrid algorithm framework, comprising the following steps:

[0045] 1) Divide the low voltage distribution area user cluster G = {G1, G2, ..., G j};

[0046] 2) Calculate the total active power variation sequence of each user cluster G' = {G1', G'2, ..., G' j};

[0047] 3) Based on the change sequence C i , calculate the average absolute variability MAV of each user cluster;

[0048] 4) Based on the mean absolute variability MAV, the user clusters are sorted in descending order, and the low-voltage distribution station area user cluster is updated, which is recorded as G'={G1',G'2,...,G' j};

[0049] 5) Based on user cluster G' g , calculate the active power change sequence C of each phase of the transformer T,k ;

[0050] 6) Calculate user cluster G' g The correlation r between the total active power change and the active power change sequence of each phase of the transformer; if the correlation is the largest and the active power of the transformer phase is greater than the user cluster G' g The total active power is the current user cluster G' g Phase sequence; the initial value of g is 1;

[0051] 7) Determine whether g>j is established. If not, return to step 5). If so, end the phase sequence identification and output the identification result.

[0052] In step 1), dividing the low-voltage distribution station area into user clusters means dividing the low-voltage distribution station area into user clusters based on the principle of voltage similarity.

[0053] The voltage similarity principle means that the voltage fluctuations of users belonging to the same phase sequence show similarity.

[0054] In step 1), the step of dividing the low-voltage distribution station area user clusters includes:

[0055] 1.1) Cluster the n users in the low-voltage distribution area into j user clusters, denoted by G = {G1, G2, …, G j};

[0056] 1.2) Classify the zero-power user into the user cluster of non-zero-power users whose voltage fluctuation similarity with the zero-power user is greater than a certain value (first divide the users with non-zero active power into j user clusters using the PAM clustering algorithm based on the minimum Davis-Bourdin index, and then classify the corresponding users with zero active power into these j user clusters), and update the user cluster G.

[0057] In step 1.1), the method for clustering n users in the low-voltage distribution station area into j user clusters includes a PAM clustering algorithm based on the minimum Davis-Boulting index.

[0058] In step 2), the total active power change sequence of each user cluster is recorded as

[0059] The total active power change of each user cluster is as follows:

[0060]

[0061] In the formula, is the total active power of user i at different times.

[0062] In step 3), the mean absolute variability MAV is as follows:

[0063]

[0064] Where M is the length of the active sequence.

[0065] In step 5), the active power change sequence of each phase of the transformer is recorded as

[0066] Among them, the active power change c of the transformer k phase t (P T,k ) is as follows:

[0067]

[0068] Where, t = 1, 2, ..., T M ; T M is the number of sample sequences between two different moments; is the active power of the transformer k phase; P Loss,k is the line loss of the transformer k phase; h i Represents user cluster G' g The decision variable of whether user i belongs to phase sequence k; k = {A, B, C} represents the phase sequence; n is the number of users in the current low-voltage distribution station area; for example, assuming that m = 5 and t = 3 at this time, then formula 3 calculates p 5 -p 2 Active power changes at these two moments.

[0069] In step 6), the correlation r is as follows:

[0070]

[0071] In the formula, For user cluster G' g The mean of the total active power change sequence and the active power change sequence of each phase of the transformer.

[0072] Embodiment 2:

[0073] A method for identifying phase sequence of users in a low-voltage distribution area based on a hybrid algorithm framework comprises the following steps:

[0074] 1) Divide the low voltage distribution area user cluster G = {G1, G2, ..., G j};

[0075] 2) Calculate the total active power variation sequence of each user cluster G' = {G1', G'2, ..., G' j};

[0076] 3) Based on the change sequence C i , calculate the average absolute variability MAV of each user cluster;

[0077] 4) Based on the mean absolute variability MAV, the user clusters are sorted in descending order, and the low-voltage distribution station area user cluster is updated, which is recorded as G'={G1',G'2,...,G' j};

[0078] 5) Based on user cluster G' g , calculate the active power change sequence C of each phase of the transformer T,k ;

[0079] 6) Calculate user cluster G' g The correlation r between the total active power change and the active power change sequence of each phase of the transformer; if the correlation is the largest and the active power of the transformer phase is greater than the user cluster G' g The total active power is the current user cluster G' g Phase sequence; the initial value of g is 1;

[0080] 7) Determine whether g>j is established. If not, return to step 5). If so, end the phase sequence identification and output the identification result.

[0081] Embodiment 3:

[0082] A method for identifying phase sequence of users in a low-voltage distribution station area based on a hybrid algorithm framework, the technical content of which is the same as that of Example 2. Furthermore, in step 1), dividing the low-voltage distribution station area user clusters means dividing the low-voltage distribution station area user clusters based on the principle of voltage similarity.

[0083] Embodiment 4:

[0084] A method for identifying the phase sequence of users in a low-voltage distribution station area based on a hybrid algorithm framework, the technical content of which is the same as any one of Examples 2-3. Furthermore, the voltage similarity principle means that the voltage fluctuations of users belonging to the same phase sequence show similarity.

[0085] Embodiment 5:

[0086] A method for identifying phase sequence of users in a low-voltage distribution station area based on a hybrid algorithm framework, the technical content of which is the same as any one of embodiments 2-4, further, in step 1), the step of dividing the user clusters in the low-voltage distribution station area includes:

[0087] 1.1) Cluster the n users in the low-voltage distribution area into j user clusters, denoted by G = {G1, G2, …, G j};

[0088] 1.2) The zero-power user is classified into the user cluster where the non-zero-power users whose voltage fluctuation similarity with the zero-power user is greater than a certain value are located, and the user cluster G is updated.

[0089] Embodiment 6:

[0090] A method for identifying the phase sequence of users in a low-voltage distribution station area based on a hybrid algorithm framework, the technical content of which is the same as any one of embodiments 2-5. Furthermore, in step 1.1), the method for clustering n users in the low-voltage distribution station area into j user clusters includes a PAM clustering algorithm based on the minimum Davis-Boulting index.

[0091] Embodiment 7:

[0092] A method for identifying phase sequence of users in a low-voltage distribution station area based on a hybrid algorithm framework, the technical content of which is the same as any one of Embodiments 2-6. Further, in step 2), the total active power change sequence under each user cluster is recorded as

[0093] The total active power change of each user cluster is as follows:

[0094]

[0095] In the formula, is the total active power of user i at time m.

[0096] Embodiment 8:

[0097] A method for identifying phase sequence of users in a low-voltage distribution station area based on a hybrid algorithm framework, the technical content of which is the same as any one of Embodiments 2-7, and further, in step 3), the mean absolute variability MAV is as follows:

[0098]

[0099] Where M is the length of the active sequence.

[0100] Embodiment 9:

[0101] A method for identifying phase sequence of users in a low-voltage distribution area based on a hybrid algorithm framework, the technical content of which is the same as any one of Embodiments 2-8, and further, in step 5), the active power change sequence of each phase of the transformer is recorded as

[0102] Among them, the active power change c of the transformer k phase t (P T,k ) is as follows:

[0103]

[0104] Where, t = 1, 2, ..., T M ; T M is the number of sample sequences between two different moments; is the active power of the transformer k phase; P Loss,k is the line loss of the transformer k phase; h i Represents user cluster G' g The decision variable of whether user i belongs to phase sequence k; k = {A, B, C} represents the phase sequence; n is the number of users in the current low-voltage distribution station area; m represents the time.

[0105] Embodiment 10:

[0106] A method for identifying phase sequence of users in a low-voltage distribution station area based on a hybrid algorithm framework, the technical content of which is the same as any one of embodiments 2-9, and further, in step 6), the correlation r is as follows:

[0107]

[0108] In the formula, For user cluster G' g The mean of the total active power change sequence and the active power change sequence of each phase of the transformer.

[0109] Embodiment 11:

[0110] A method for identifying the phase sequence of users in a low-voltage distribution area based on a hybrid algorithm framework is proposed. The method is based on the principle of voltage fluctuation similarity. In a low-voltage distribution area, due to factors such as power flow, load distribution, and line impedance, the voltage fluctuations of users with the same phase sequence show a high degree of similarity. The similarity of voltage fluctuations is more significant between users with a close electrical distance.

[0111] Figure 1 It represents the voltage data of users with different phase sequences in a real low-voltage distribution station at 24 o'clock on a certain day. It can be seen that users belonging to the three phases A, B, and C have a significant similar trend of voltage fluctuation. Figure 2 The voltage fluctuation curve of six users in phase B at 24 o'clock on a certain day is shown in the figure. It can be seen from the figure that even among users in the same phase sequence, users 1-2 and users 3-6 show a higher similarity in voltage fluctuation. Combined with the physical connection of the users on site, the topological locations of these two groups of users are closer, which leads to an increase in the similarity of their voltage fluctuations.

[0112] In addition, according to the power conservation theorem, the active power generated by each phase of the low-voltage distribution transformer is equal to the sum of the active power consumed by all users under this phase and the line loss:

[0113]

[0114] Where k = {A, B, C} represents the phase sequence, P T,k is the active power of a phase of the transformer, n is the number of users in the area, P Loss,k is the line loss of phase k, h i The decision variable indicating whether user i belongs to phase sequence k exists:

[0115]

[0116] Formula (1) shows that there is a direct approximate linear correlation between the active power of a phase of the distribution transformer and the active power of the user under this phase. Although the active power loss under the corresponding phase sequence will weaken the correlation to a certain extent, it can still be used as a basis for judging the phase sequence of the user.

[0117] Since the trend of power change over time can better reveal whether the power consumption pattern of the user and the phase to which they belong is similar than the absolute value of power, the relationship between them can be inferred. Therefore, the change of active power sequence of each phase of the transformer and the user can be expressed as:

[0118]

[0119] Among them, m represents the time, t is the number of samples between two different time points. According to the literature, in this patent, t = {1, 2, 3, 4}, the sequence of active power changes is obtained as follows:

[0120] C=[c1,c2,c3,c4] (5)

[0121] In addition, formula (6) represents the Pearson correlation coefficient, which is used to determine the correlation between the active power changes of each phase of the user and the transformer. i , C T , k represent the active power change sequence of user i and transformer k phase respectively.

[0122]

[0123] like Figure 2 FIG. 1 is a flow chart of a method for identifying a phase sequence of a user in a low-voltage distribution station area based on a hybrid algorithm framework of the present invention, comprising the following steps:

[0124] Step 201, based on the principle of voltage similarity, the n users in the area are clustered into j user clusters G using the Partitioning Around Medoids (PAM) clustering algorithm based on the minimum Davies-Bouldin (DB) index, where G = {G1, G2, ..., G jAt the same time, since some users in low-voltage distribution areas do not use electricity (mostly common in rural distribution areas), it is impossible to calculate the correlation sequence of active power changes. Therefore, while considering the minimum DB index, it is necessary to classify zero-power users into appropriate user clusters to ensure that the phase sequence of these users can also be effectively identified. Use equations (3) and (5) to calculate the active power correlation sequence of each phase of the transformer, and use equations (4) and (5) to calculate the correlation sequence C of the total active power under each user cluster. Subsequently, use equation (7) to calculate the mean absolute variability (MAV) of each user cluster to sort the recognition order of the user clusters. The larger the MAV, the more obvious the power change, which can reduce the difficulty of identification and increase the accuracy of subsequent user identification. Among them, T M =4, M is the length of the active sequence.

[0125]

[0126] Step 202: The larger the MAV of the user cluster Gj is, the higher the recognition order is. Calculate r(C j ,C T ,k), select the phase sequence with the largest correlation and the sum of the power provided by the phase transformer minus the active power of the user cluster is greater than 0, which is the phase sequence to which the current user cluster belongs.

[0127] Step 203: determine whether all user clusters have been identified. If not, update C. T,k Otherwise, the phase sequence identification is terminated and the identification result is output.

Claims

1. A method for identifying phase sequence of users in low-voltage distribution area based on a hybrid algorithm framework, characterized in that: The following steps are involved: 1) Divide the low voltage distribution area user cluster G = {G1, G2, ..., G j }. 2) Calculate the total active power variation sequence of each user cluster G' = {G'1, G'2, ..., G' j }; 3) Based on the change sequence C i , calculate the average absolute variability MAV of each user cluster; 4) Based on the mean absolute variability MAV, the user clusters are sorted in descending order, and the low-voltage distribution station area user cluster is updated, which is recorded as G'={G'1,G'2,...,G' j }; 5) Based on user cluster G' g , calculate the active power change sequence C of each phase of the transformer T,k ; 6) Calculate user cluster G' g The correlation r between the total active power change and the active power change sequence of each phase of the transformer; if the correlation is the largest and the active power of the transformer phase is greater than the user cluster G' g The total active power is the current user cluster G' g Phase sequence; the initial value of g is 1; 7) Determine whether g>j is established. If not, return to step 5). If so, end the phase sequence identification and output the identification result.

2. According to claim 1, a method for identifying phase sequence of users in low-voltage distribution station area based on a hybrid algorithm framework is characterized in that: In step 1), dividing the low-voltage distribution station area into user clusters means dividing the low-voltage distribution station area into user clusters based on the principle of voltage similarity.

3. According to claim 2, a method for identifying phase sequence of users in low-voltage distribution station area based on a hybrid algorithm framework is characterized in that: The voltage similarity principle means that the voltage fluctuations of users belonging to the same phase sequence show similarity.

4. According to claim 1, a method for identifying phase sequence of users in low-voltage distribution station area based on a hybrid algorithm framework is characterized in that: In step 1), the step of dividing the low-voltage distribution station area user clusters includes: 1.1) Cluster the n users in the low-voltage distribution area into j user clusters, denoted by G = {G1, G2, …, G j }; 1.2) The zero-power user is classified into the user cluster of non-zero-power users whose voltage fluctuation similarity with the zero-power user is greater than a certain value, and the user cluster G is updated.

5. According to claim 4, a method for identifying phase sequence of users in low-voltage distribution area based on a hybrid algorithm framework is characterized in that: In step 1.1), the method for clustering n users in the low-voltage distribution station area into j user clusters includes a PAM clustering algorithm based on the minimum Davis-Boulting index.

6. According to claim 1, a method for identifying phase sequence of users in low-voltage distribution station area based on a hybrid algorithm framework is characterized in that: In step 2), the total active power change sequence of each user cluster is recorded as The total active power change of each user cluster is as follows: In the formula, is the total active power of user i at different times.

7. The method for identifying phase sequence of users in low-voltage distribution area based on a hybrid algorithm framework according to claim 1 is characterized in that: In step 3), the mean absolute variability MAV is as follows: Where M is the length of the active sequence.

8. The method for identifying phase sequence of users in low-voltage distribution area based on a hybrid algorithm framework according to claim 1 is characterized in that: In step 5), the active power change sequence of each phase of the transformer is recorded as Among them, the active power change c of the transformer k phase t (P T,k ) is as follows: Where, t = 1, 2, ..., T M ; T M is the number of sample sequences between two different moments; is the active power of the transformer k phase; P Loss,k is the line loss of the transformer k phase; h i Represents user cluster G' g The decision variable of whether user i belongs to phase sequence k; k = {A, B, C} represents the phase sequence; n is the number of users in the current low-voltage distribution station area; m represents the time.

9. The method for identifying phase sequence of users in low-voltage distribution area based on a hybrid algorithm framework according to claim 1 is characterized in that: In step 6), the correlation r is as follows: In the formula, For user cluster G' g The mean of the total active power change sequence and the active power change sequence of each phase of the transformer.

Images