Method and system for evaluating state of power distribution network based on small sample state

Abstract

The embodiment of the invention provides a power distribution network state evaluation method and system based on a small sample state, and belongs to the technical field of power distribution network data maintenance. The evaluation method comprises the following steps: constructing a matrix topology model and a graph topology model for power grid multi-source data fusion; acquiring power grid data; adding the power grid data into the matrix topology model to obtain a power distribution network state model; carrying out checking operation, correction operation and analysis operation on the power distribution network state model in sequence; adding the power grid data into the graph topology model, and executing balance analysis operation so as to distribute power grid graph topology; performing data completion operation on the power distribution network state model after the balance analysis operation; and encoding the power distribution network state model and the power distribution network graph topology after the data completion operation, and inputting the encoded power distribution network state model and the encoded power distribution network graph topology into a preset power distribution network state estimation network to obtain an evaluation result.

Description

Technical Field

[0001] The present invention relates to the technical field of maintenance of distribution network data, and in particular to a method, device, storage medium and processor for processing field information. Background Art

[0002] Grid data generally refers to data generated from all aspects of the grid (such as power generation, transmission, transformation, distribution and consumption), mainly including data from energy management systems (EMS), distribution network management systems, wide-area measurement management systems, production management systems, grid dispatching management systems and fault management systems. These data involve the parameters of grid equipment, the topological relationship of equipment during operation and measurement data, which are crucial for monitoring and analyzing grid operations.

[0003] In order to evaluate the overall state of the power grid, a neural network model can be used to train and evaluate the power grid data. However, due to the complex types of power grid data and the huge parameters of each electrical device, it is difficult to achieve complete alignment of the state when integrating the data. Therefore, it is currently difficult to implement the complete training process of applying the neural network model to power grid data. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to provide a method and system for evaluating the state of a distribution network based on a small sample state, wherein the method and system can achieve accurate evaluation of power grid data.

[0005] In order to achieve the above object, an embodiment of the present invention provides a method for evaluating the state of a distribution network based on a small sample state, comprising:

[0006] Construct matrix topology model and graph topology model for multi-source data fusion of power grid;

[0007] Obtain power grid data;

[0008] Adding the power grid data to the matrix topology model to obtain a distribution network state model;

[0009] Performing verification, correction and analysis operations on the distribution network state model in sequence;

[0010] adding the power grid data to the graph topology model and performing a balance analysis operation to distribute the power grid graph topology;

[0011] Performing a data completion operation on the distribution network state model after the balance analysis operation;

[0012] The distribution network state model and distribution network diagram topology after the data completion operation are encoded and input into a preset distribution network state estimation network to obtain an evaluation result.

[0013] Optionally, a matrix topology model and a graph topology model for fusion of multi-source data of the power grid are constructed, including:

[0014] Obtaining a correlation matrix table of a preset power grid topology to obtain a sparse matrix;

[0015] The sparse matrix is ​​processed by using a Mapreduce method to obtain a power grid topology node analysis result;

[0016] The matrix topology model is determined according to the power grid topology node analysis result.

[0017] Optionally, a matrix topology model and a graph topology model for fusion of multi-source data of the power grid are constructed, including:

[0018] Parse the distribution of power grid data into a database-oriented node and edge data set to obtain a migration data set;

[0019] Performing a search and analysis operation on the graph database of the migration data set to obtain a local model of the connected region;

[0020] Superimposing the disconnection device status on the local model of the connected area to obtain an electrical island;

[0021] The graph topology model is constructed with each of the electrical islands as a boundary.

[0022] Optionally, the distribution network state model is sequentially subjected to a verification operation, a correction operation and an analysis operation, including:

[0023] For each indicator in the distribution network state model, construct an indicator abnormal threshold based on corresponding historical data;

[0024] Determine the state of the indicator at the corresponding sampling time according to the indicator abnormality threshold;

[0025] Using a time window sliding method to check the sampling time of the indicator according to the state;

[0026] Adopting an export voltage normalization algorithm to correct the overall sampling clock of the indicator;

[0027] A virtual agent method is used to analyze the status of the line where the indicator is located based on the line voltage distribution principle.

[0028] Optionally, for each indicator in the distribution network state model, an indicator abnormality threshold is constructed based on corresponding historical data, including:

[0029] Determine a state distribution of a state value of the indicator based on the historical data;

[0030] The indicator abnormality threshold is determined according to the state distribution.

[0031] Optionally, a time window sliding method is used to check the sampling time of the indicator according to the state, including:

[0032] Determine historical status in power grid data;

[0033] Comparing the historical state with the state to determine whether the sampling time corresponding to the state is correct;

[0034] When the historical state and the state are the same, determining that the sampling time is correct;

[0035] In the case where the historical state and the state are different, searching for a state that is the same as the historical state around the sampling time, and determining whether a state that is the same as the historical state can be found;

[0036] If a state identical to the historical state cannot be found, deleting the sampling moment;

[0037] When a state identical to the historical state can be found, the sampling time corresponding to the found state is corrected to the current sampling time.

[0038] Optionally, using an exit voltage normalization algorithm to correct the overall sampling clock of the indicator includes:

[0039] Taking each node as the center, determine the voltage variation curve of the corresponding three-phase line;

[0040] Determining whether the three-phase load of the node complies with the three-phase load principle;

[0041] When it is determined that the three-phase load of the node does not conform to the three-phase load principle, the voltage variation curve is moved forward or backward to correct the sampling clock.

[0042] Optionally, a virtual agent method is used to analyze and mark the state of the line where the indicator is located according to the line voltage distribution principle, including:

[0043] A virtual agent method is used to determine whether the change in the state value of the indicator on each line satisfies the line voltage distribution principle;

[0044] In the case where the change in the state value of the indicator on the line does not satisfy the line voltage distribution principle, the state is determined to be abnormal.

[0045] Optionally, encoding the distribution network state model and distribution network graph topology after the data completion operation includes:

[0046] Using a preset parameter pre-training model to generate a corresponding first feature matrix according to the distribution network state model;

[0047] Using a preset graph pre-training model to generate a corresponding second feature matrix according to the distribution network graph topology;

[0048] The first feature matrix and the second feature matrix are combined by a feature integration unit to obtain an encoding result.

[0049] On the other hand, the present invention further provides an evaluation system for a distribution network state based on a small sample state, wherein the evaluation system comprises a processor, and the processor is configured to execute any of the evaluation methods described above.

[0050] Through the above technical solution, the embodiment of the present invention provides an evaluation method and system for the state of the distribution network based on a small sample state. The evaluation method and system process the power grid data by constructing a matrix topology model and a graph topology model for the fusion of multi-source data of the power grid, and combine the verification, correction and analysis operations for different data types to achieve the integration of the power grid data. Finally, the overall state of the power grid is evaluated through the distribution network state estimation network. Compared with the prior art, the evaluation method and system provided by the present invention realizes the training and evaluation of power grid data because it combines different processing methods for different data.

[0051] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0053] Figure 1 is a flow chart of a method for evaluating a distribution network state based on a small sample state according to an embodiment of the present invention;

[0054] Figure 2 is a flow chart of a verification operation, a correction operation and an analysis operation based on a small sample state according to an embodiment of the present invention;

[0055] Figure 3 is a flow chart of a time window sliding method according to one embodiment of the present invention;

[0056] Figure 4 is a schematic diagram of a three-phase load principle according to an embodiment of the present invention;

[0057] Figure 5 It is a schematic diagram of the connection relationship between a parameter pre-training model, a graph pre-training model and a feature integration unit according to an embodiment of the present invention. DETAILED DESCRIPTION

[0058] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0059] like Figure 1 FIG. 1 is a flow chart of a method for evaluating a distribution network state based on a small sample state according to an embodiment of the present invention. Figure 1 In the present invention, the evaluation method may include the following steps:

[0060] In step S10, a matrix topology model and a graph topology model for fusion of multi-source data of the power grid are constructed;

[0061] In step S11, power grid data is obtained;

[0062] In step S12, the power grid data is added to the matrix topology model to obtain a distribution network state model;

[0063] In step S13, the distribution network state model is sequentially subjected to verification, correction and analysis operations;

[0064] In step S14, the power grid data is added to the graph topology model, and a balance analysis operation is performed to distribute the power grid graph topology;

[0065] In step S15, a data completion operation is performed on the distribution network state model after the balance analysis operation;

[0066] In step S16, the distribution network state model and the distribution network diagram topology after the data completion operation are encoded and input into a preset distribution network state estimation network to obtain an evaluation result.

[0067] In this Figure 1 In the evaluation method shown, step S10 can be used to construct a matrix topology model and a graph topology model for fusion of multi-source data of a power grid. The matrix topology model can represent the result of standardization of power grid data. Specifically, in one example of the present invention, the method for constructing the matrix topology model can be to first obtain an association matrix table of a preset power grid topology to obtain a sparse matrix. Then, the sparse matrix is ​​processed using a Mapreduce method to obtain a power grid topology node analysis result. Finally, the topology model is determined based on the power grid topology node analysis result.

[0068] The graph topology model can be used to represent the association relationship of the images of electrical equipment. Specifically, in one example of the present invention, the method for obtaining the graph topology model can be to first parse the distribution of power grid data into a database-oriented node and edge data set form to obtain a migration data set. Then, the graph database of the migration data set is searched and analyzed to obtain a local model of the connected area. The disconnection device status is then superimposed on the local model of the connected area to obtain an electrical island. Finally, the graph topology model is constructed with each electrical island as the boundary. Among them, the step of parsing the distribution of power grid data into a node and edge data set form is mainly due to the fact that the power grid data itself has many types and is difficult to process directly. Through the expression of nodes and edges, the local "islanding" of the power grid data is achieved, which facilitates subsequent processing. The search and analysis operation is to classify the images of electrical equipment according to the determined nodes and edges, thereby realizing the process of numbering or coding the electrical equipment in each electrical island.

[0069] Step S13 may be used to sequentially perform a calibration operation, a correction operation, and an analysis operation on the distribution network state model. The specific processes of the calibration operation, the correction operation, and the analysis operation may be in various forms known to those skilled in the art. In one example of the present invention, the specific method of step S13 may include: Figure 2 In the steps shown in Figure 2 In the step S13, the step S13 may include the following steps:

[0070] In step S20, for each indicator in the distribution network state model, an indicator abnormality threshold is constructed based on the corresponding historical data;

[0071] In step S21, the state of the indicator at the corresponding sampling time is determined according to the indicator abnormality threshold;

[0072] In step S22, the sampling time of the indicator is checked according to the state by using the time window sliding method;

[0073] In step S23, the overall sampling clock of the indicator is corrected using the outlet voltage normalization algorithm;

[0074] In step S24, a virtual agent method is used to analyze the state of the line where the marking indicator is located according to the line voltage distribution principle.

[0075] In this Figure 2In the method shown, step S20 can be used to construct an indicator abnormality threshold for each indicator in the distribution network state model based on the corresponding historical data. For the indicator abnormality threshold, it can be a variety of forms known to those skilled in the art. In one example of the present invention, the indicator abnormality threshold can be manually set by a staff member. In another example of the present invention, the indicator abnormality threshold can be a state distribution of the state value of the indicator first determined based on historical data, and then the indicator abnormality threshold is determined based on the state distribution. Further, the state distribution can be based on a curve with the horizontal axis as the sampling moment and the vertical axis as the state value, and the indicator abnormality threshold can be calculated from the maximum value of the state value, and the cumulative number is greater than or equal to 2% of the total number of state values. After determining the indicator abnormality threshold, it can be further used to determine the state of the indicator at the corresponding sampling moment according to the indicator abnormality threshold through step S21.

[0076] Step S22 may be to use a time window sliding method to check the sampling time of the indicator according to the state. Specifically, the time window sliding method may include: Figure 3 In the steps shown in Figure 3 In the time window sliding method, the time window sliding method may include the following steps:

[0077] In step S30, determining the historical status in the power grid data;

[0078] In step S31, the historical state is compared with the state to determine whether the sampling time corresponding to the state is correct;

[0079] In step S32, when the historical state and the state are the same, it is determined that the sampling time is correct;

[0080] In step S33, when the historical state and the current state are different, a state identical to the historical state is searched around the sampling time, and it is determined whether the state identical to the historical state can be found.

[0081] In step S34, if the same state as the historical state cannot be found, the sampling time is deleted;

[0082] In step S35, when a state identical to the historical state can be found, the sampling time corresponding to the found state is corrected to the current sampling time.

[0083] Step S23 may be to use the export voltage normalization algorithm to correct the overall sampling clock of the index. Specifically, in this example, step S23 may first determine the voltage change curve of the corresponding three-phase line with each node as the center, and then determine whether the three-phase load of the node conforms to the three-phase load principle, and if it is determined that the three-phase load of the node does not conform to the three-phase load principle, the voltage change curve is moved forward or backward to correct the sampling clock. Among them, the three-phase load principle can be as follows Figure 4 The schematic diagram shown. Specifically, the three-phase load principle is generally applied to the outlet of the distribution transformer (distribution transformer outlet). The main reason for the asymmetry of the three-phase voltage at the outlet of the distribution transformer is that the distribution transformer winding in the distribution network has two connection methods: Dyn11 and yyn0. Since the high-voltage side of the distribution transformer with the yyn0 connection method is a star connection, and the neutral point of the high-voltage side winding is not grounded, when the three-phase load on the low-voltage side is unbalanced, the zero-sequence current cannot flow in the high-voltage side winding, and the zero-sequence current will become the excitation current, generating a zero-sequence potential in the high-voltage side winding, causing the neutral point of the high-voltage side to shift, and the three-phase voltage on the high-voltage side will be asymmetric, and the corresponding three-phase voltage at the outlet of the low-voltage side will also be asymmetric. Actual operating experience shows that there are many non-energy-saving transformers in rural distribution networks, and their winding connection forms are mostly yyn0 connections. Due to the influence of the unbalanced three-phase load, the voltages at these distribution transformer outlets are more or less asymmetric.

[0084] For the distribution transformer with winding connection form yyn0, since the neutral point on the high-voltage side of the distribution transformer is not grounded, when the three-phase load is balanced, the neutral point N on the high-voltage side of the distribution transformer does not shift, and the three-phase voltage at the distribution transformer outlet is symmetrical; when the three-phase load is unbalanced, the neutral point N on the high-voltage side of the distribution transformer shifts to point N′, resulting in unbalanced three-phase voltage at the distribution transformer outlet.

[0085] Step S24 may be used to analyze the state of the line where the marked indicator is located according to the line voltage distribution principle using a virtual agent method. Specifically, step S24 may be to use a virtual agent method to determine whether the change in the state value of the indicator on each line meets the line voltage distribution principle. If the change in the state value of the indicator on the line does not meet the line voltage distribution principle, the state is determined to be abnormal. Otherwise, the state is determined to be normal.

[0086] Step S14 may be used to add the power grid data into the graph topology model and perform a balance analysis operation to distribute the power grid graph topology.

[0087] Step S15 can be used to perform a data completion operation on the distribution network state model after the balance analysis operation. Specifically, step S15 can be based on the distribution network state model. The specific method of the completion operation can be in various forms known to those skilled in the art. In this embodiment, different completion operations can be used for data in different forms. In one example of the present invention, when the missing point is a single sampling point, the sampling point can be completed by a linear value method, that is, the following formula (1),

[0088]

[0089] Among them, x i is the missing sampling point, x i-1 is the sampling point to the left of the missing sampling point, x i+1 is the sampling point to the right of the missing sampling point.

[0090] In another example of the present invention, when there are multiple sampling points missing, the sampling points can be supplemented by sampling historical data. Specifically, the method can be to compare the current segment (preset sampling time period) with the sampling time period with the same state and time in the historical data. If the similarity of other sampling points around the actual sampling point meets the preset conditions, the historical data can be directly used to supplement the actual sampling point.

[0091] In step S16, the distribution network state model and distribution network graph topology after the data completion operation are encoded and input into the preset distribution network state estimation network to obtain the evaluation result. Considering that the parameter form of the distribution network model and the parameter form of the distribution network graph topology are different, the parameter form of the distribution network model is in the form of a matrix code, while the distribution network graph topology is in the form of a connected graph. Therefore, in step S16, a parameter pre-training model and a graph pre-training model can be further constructed to process the distribution network state model and the distribution network graph topology, and finally the extracted features are combined by constructing a feature integration unit to realize the encoding operation. The connection relationship is as follows: Figure 5 Specifically, step S16 may be to first use a preset parameter pre-training model to generate a corresponding first feature matrix according to the distribution network state model, then use a preset graph pre-training model to generate a corresponding second feature matrix according to the distribution network graph topology, and finally use a feature integration unit to combine the first feature matrix and the second feature matrix to obtain an encoding result.

[0092] On the other hand, the present invention also provides an evaluation system based on the state of the distribution network under a small sample state, the evaluation system comprising a processor, the processor being configured to execute any of the evaluation methods described above. Specifically, the evaluation method may include the following steps:

[0093] In step S10, a matrix topology model and a graph topology model for fusion of multi-source data of the power grid are constructed;

[0094] In step S11, power grid data is obtained;

[0095] In step S12, the power grid data is added to the matrix topology model to obtain a distribution network state model;

[0096] In step S13, the distribution network state model is sequentially subjected to verification, correction and analysis operations;

[0097] In step S14, the power grid data is added to the graph topology model, and a balance analysis operation is performed to distribute the power grid graph topology;

[0098] In step S15, a data completion operation is performed on the distribution network state model after the balance analysis operation;

[0099] In step S16, the distribution network state model and the distribution network diagram topology after the data completion operation are encoded and input into a preset distribution network state estimation network to obtain an evaluation result.

[0100] In this Figure 1 In the evaluation method shown, step S10 can be used to construct a matrix topology model and a graph topology model for fusion of multi-source data of a power grid. The matrix topology model can represent the result of standardization of power grid data. Specifically, in one example of the present invention, the method for constructing the matrix topology model can be to first obtain an association matrix table of a preset power grid topology to obtain a sparse matrix. Then, the sparse matrix is ​​processed using a Mapreduce method to obtain a power grid topology node analysis result. Finally, the topology model is determined based on the power grid topology node analysis result.

[0101] The graph topology model can be used to represent the association relationship of the images of electrical equipment. Specifically, in one example of the present invention, the method for obtaining the graph topology model can be to first parse the distribution of power grid data into a database-oriented node and edge data set form to obtain a migration data set. Then, the graph database of the migration data set is searched and analyzed to obtain a local model of the connected area. The disconnection device status is then superimposed on the local model of the connected area to obtain an electrical island. Finally, the graph topology model is constructed with each electrical island as the boundary. Among them, the step of parsing the distribution of power grid data into a node and edge data set form is mainly due to the fact that the power grid data itself has many types and is difficult to process directly. Through the expression of nodes and edges, the local "islanding" of the power grid data is achieved, which facilitates subsequent processing. The search and analysis operation is to classify the images of electrical equipment according to the determined nodes and edges, thereby realizing the process of numbering or coding the electrical equipment in each electrical island.

[0102] Step S13 may be used to sequentially perform a calibration operation, a correction operation, and an analysis operation on the distribution network state model. The specific processes of the calibration operation, the correction operation, and the analysis operation may be in various forms known to those skilled in the art. In one example of the present invention, the specific method of step S13 may include: Figure 2 In the steps shown in Figure 2 In the step S13, the step S13 may include the following steps:

[0103] In step S20, for each indicator in the distribution network state model, an indicator abnormality threshold is constructed based on the corresponding historical data;

[0104] In step S21, the state of the indicator at the corresponding sampling time is determined according to the indicator abnormality threshold;

[0105] In step S22, the sampling time of the indicator is checked according to the state by using the time window sliding method;

[0106] In step S23, the overall sampling clock of the indicator is corrected using the outlet voltage normalization algorithm;

[0107] In step S24, a virtual agent method is used to analyze the state of the line where the marking indicator is located according to the line voltage distribution principle.

[0108] In this Figure 2In the method shown, step S20 can be used to construct an indicator abnormality threshold for each indicator in the distribution network state model based on the corresponding historical data. For the indicator abnormality threshold, it can be a variety of forms known to those skilled in the art. In one example of the present invention, the indicator abnormality threshold can be manually set by a staff member. In another example of the present invention, the indicator abnormality threshold can be a state distribution of the state value of the indicator first determined based on historical data, and then the indicator abnormality threshold is determined based on the state distribution. Further, the state distribution can be based on a curve with the horizontal axis as the sampling moment and the vertical axis as the state value, and the indicator abnormality threshold can be calculated from the maximum value of the state value, and the cumulative number is greater than or equal to 2% of the total number of state values. After determining the indicator abnormality threshold, it can be further used to determine the state of the indicator at the corresponding sampling moment according to the indicator abnormality threshold through step S21.

[0109] Step S22 may be to use a time window sliding method to check the sampling time of the indicator according to the state. Specifically, the time window sliding method may include: Figure 3 In the steps shown in Figure 3 In the time window sliding method, the time window sliding method may include the following steps:

[0110] In step S30, determining the historical status in the power grid data;

[0111] In step S31, the historical state is compared with the state to determine whether the sampling time corresponding to the state is correct;

[0112] In step S32, when the historical state and the state are the same, it is determined that the sampling time is correct;

[0113] In step S33, when the historical state and the current state are different, a state identical to the historical state is searched around the sampling time, and it is determined whether the state identical to the historical state can be found.

[0114] In step S34, if the same state as the historical state cannot be found, the sampling time is deleted;

[0115] In step S35, when a state identical to the historical state can be found, the sampling time corresponding to the found state is corrected to the current sampling time.

[0116] Step S23 may be to use the export voltage normalization algorithm to correct the overall sampling clock of the index. Specifically, in this example, step S23 may first determine the voltage change curve of the corresponding three-phase line with each node as the center, and then determine whether the three-phase load of the node conforms to the three-phase load principle, and if it is determined that the three-phase load of the node does not conform to the three-phase load principle, the voltage change curve is moved forward or backward to correct the sampling clock. Among them, the three-phase load principle can be as follows Figure 4 The schematic diagram shown. Specifically, the three-phase load principle is generally applied to the outlet of the distribution transformer (distribution transformer outlet). The main reason for the asymmetry of the three-phase voltage at the outlet of the distribution transformer is that the distribution transformer winding in the distribution network has two connection methods: Dyn11 and yyn0. Since the high-voltage side of the distribution transformer with the yyn0 connection method is a star connection, and the neutral point of the high-voltage side winding is not grounded, when the three-phase load on the low-voltage side is unbalanced, the zero-sequence current cannot flow in the high-voltage side winding, and the zero-sequence current will become the excitation current, generating a zero-sequence potential in the high-voltage side winding, causing the neutral point of the high-voltage side to shift, and the three-phase voltage on the high-voltage side will be asymmetric, and the corresponding three-phase voltage at the outlet of the low-voltage side will also be asymmetric. Actual operating experience shows that there are many non-energy-saving transformers in rural distribution networks, and their winding connection forms are mostly yyn0 connections. Due to the influence of the unbalanced three-phase load, the voltages at these distribution transformer outlets are more or less asymmetric.

[0117] For the distribution transformer with winding connection form yyn0, since the neutral point on the high-voltage side of the distribution transformer is not grounded, when the three-phase load is balanced, the neutral point N on the high-voltage side of the distribution transformer does not shift, and the three-phase voltage at the distribution transformer outlet is symmetrical; when the three-phase load is unbalanced, the neutral point N on the high-voltage side of the distribution transformer shifts to point N′, resulting in unbalanced three-phase voltage at the distribution transformer outlet.

[0118] Step S24 may be used to analyze the state of the line where the marked indicator is located according to the line voltage distribution principle using a virtual agent method. Specifically, step S24 may be to use a virtual agent method to determine whether the change in the state value of the indicator on each line meets the line voltage distribution principle. If the change in the state value of the indicator on the line does not meet the line voltage distribution principle, the state is determined to be abnormal. Otherwise, the state is determined to be normal.

[0119] Step S14 may be used to add the power grid data into the graph topology model and perform a balance analysis operation to distribute the power grid graph topology.

[0120] Step S15 can be used to perform a data completion operation on the distribution network state model after the balance analysis operation. Specifically, step S15 can be based on the distribution network state model. The specific method of the completion operation can be in various forms known to those skilled in the art. In this embodiment, different completion operations can be used for data in different forms. In one example of the present invention, when the missing point is a single sampling point, the sampling point can be completed by a linear value method, that is, the following formula (1),

[0121]

[0122] Among them, x i is the missing sampling point, x i-1 is the sampling point to the left of the missing sampling point, x i+1 is the sampling point to the right of the missing sampling point.

[0123] In another example of the present invention, when there are multiple sampling points missing, the sampling points can be supplemented by sampling historical data. Specifically, the method can be to compare the current segment (preset sampling time period) with the sampling time period with the same state and time in the historical data. If the similarity of other sampling points around the actual sampling point meets the preset conditions, the historical data can be directly used to supplement the actual sampling point.

[0124] In step S16, the distribution network state model and distribution network graph topology after the data completion operation are encoded and input into the preset distribution network state estimation network to obtain the evaluation result. Considering that the parameter form of the distribution network model and the parameter form of the distribution network graph topology are different, the parameter form of the distribution network model is in the form of a matrix code, while the distribution network graph topology is in the form of a connected graph. Therefore, in step S16, a parameter pre-training model and a graph pre-training model can be further constructed to process the distribution network state model and the distribution network graph topology, and finally the extracted features are combined by constructing a feature integration unit to realize the encoding operation. The connection relationship is as follows: Figure 5 Specifically, step S16 may be to first use a preset parameter pre-training model to generate a corresponding first feature matrix according to the distribution network state model, then use a preset graph pre-training model to generate a corresponding second feature matrix according to the distribution network graph topology, and finally use a feature integration unit to combine the first feature matrix and the second feature matrix to obtain an encoding result.

[0125] Through the above technical solution, the embodiment of the present invention provides an evaluation method and system for the state of the distribution network based on a small sample state. The evaluation method and system process the power grid data by constructing a matrix topology model and a graph topology model for the fusion of multi-source data of the power grid, and combine the verification, correction and analysis operations for different data types to achieve the integration of the power grid data. Finally, the overall state of the power grid is evaluated through the distribution network state estimation network. Compared with the prior art, the evaluation method and system provided by the present invention realizes the training and evaluation of power grid data because it combines different processing methods for different data.

[0126] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0127] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0128] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0129] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0130] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0131] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0132] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0133] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0134] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for evaluating the state of a distribution network based on a small sample state, characterized in that: The evaluation methods include: Construct matrix topology model and graph topology model for multi-source data fusion of power grid; Obtain power grid data; Adding the power grid data to the matrix topology model to obtain a distribution network state model; Performing verification, correction and analysis operations on the distribution network state model in sequence; adding the power grid data to the graph topology model and performing a balance analysis operation to distribute the power grid graph topology; Performing a data completion operation on the distribution network state model after the balance analysis operation; The distribution network state model and distribution network diagram topology after the data completion operation are encoded and input into a preset distribution network state estimation network to obtain an evaluation result.

2. The evaluation method according to claim 1, characterized in that: Construct matrix topology model and graph topology model for multi-source data fusion of power grid, including: Obtaining a correlation matrix table of a preset power grid topology to obtain a sparse matrix; The sparse matrix is ​​processed by using a Mapreduce method to obtain a power grid topology node analysis result; The matrix topology model is determined according to the power grid topology node analysis result.

3. The evaluation method according to claim 1, characterized in that: Construct matrix topology model and graph topology model for multi-source data fusion of power grid, including: Parse the distribution of power grid data into a database-oriented node and edge data set to obtain a migration data set; Performing a search and analysis operation on the graph database of the migration data set to obtain a local model of the connected region; Superimposing the disconnection device status on the local model of the connected area to obtain an electrical island; The graph topology model is constructed with each of the electrical islands as a boundary.

4. The evaluation method according to claim 1, characterized in that: The distribution network state model is sequentially subjected to verification, correction and analysis operations, including: For each indicator in the distribution network state model, construct an indicator abnormal threshold based on corresponding historical data; Determine the state of the indicator at the corresponding sampling time according to the indicator abnormality threshold; Using a time window sliding method to check the sampling time of the indicator according to the state; Adopting an export voltage normalization algorithm to correct the overall sampling clock of the indicator; A virtual agent method is used to analyze the status of the line where the indicator is located based on the line voltage distribution principle.

5. The evaluation method according to claim 4, characterized in that: For each indicator in the distribution network state model, an indicator abnormality threshold is constructed based on the corresponding historical data, including: Determine a state distribution of a state value of the indicator based on the historical data; The indicator abnormality threshold is determined according to the state distribution.

6. The evaluation method according to claim 4, characterized in that: The sampling time of the indicator is checked according to the state by using a time window sliding method, including: Determine historical status in grid data; Comparing the historical state with the state to determine whether the sampling time corresponding to the state is correct; When the historical state and the state are the same, determining that the sampling time is correct; In the case where the historical state and the state are different, searching for a state that is the same as the historical state around the sampling time, and determining whether a state that is the same as the historical state can be found; If a state identical to the historical state cannot be found, deleting the sampling moment; When a state identical to the historical state can be found, the sampling time corresponding to the found state is corrected to the current sampling time.

7. The evaluation method according to claim 4, characterized in that: The overall sampling clock of the indicator is corrected by using an export voltage normalization algorithm, including: Taking each node as the center, determine the voltage variation curve of the corresponding three-phase line; Determining whether the three-phase load of the node complies with the three-phase load principle; When it is determined that the three-phase load of the node does not conform to the three-phase load principle, the voltage variation curve is moved forward or backward to correct the sampling clock.

8. The evaluation method according to claim 4, characterized in that: The virtual agent method is used to analyze and mark the state of the line where the indicator is located according to the line voltage distribution principle, including: A virtual agent method is used to determine whether the change in the state value of the indicator on each line satisfies the line voltage distribution principle; In the case where the change in the state value of the indicator on the line does not satisfy the line voltage distribution principle, the state is determined to be abnormal.

9. The evaluation method according to claim 1, characterized in that: The distribution network state model and the distribution network diagram topology after the data completion operation are encoded, including: Using a preset parameter pre-training model to generate a corresponding first feature matrix according to the distribution network state model; Using a preset graph pre-training model to generate a corresponding second feature matrix according to the distribution network graph topology; The first feature matrix and the second feature matrix are combined by a feature integration unit to obtain an encoding result.

10. A distribution network status evaluation system based on a small sample state, characterized in that: The evaluation system comprises a processor, and the processor is configured to execute the evaluation method according to any one of claims 1 to 9.

Images