Cable power transmission data completion method and device

Abstract

The invention discloses a cable power transmission data completion method and device, and belongs to the technical field of cable data, and the method comprises the steps: obtaining a cable power transmission data sequence, power grid topological data and a cable transient timestamp; generating a topology label sequence of the cable power transmission data sequence according to the power grid topology data; generating a transient label sequence of the cable power transmission data sequence based on the cable transient time; performing feature extraction on the cable power transmission data sequence to generate a data feature sequence and a mask feature sequence; generating cable analysis data based on the topological tag sequence, the transient tag sequence, the data feature sequence and the mask feature sequence; based on the cable analysis data, performing data estimation on the cable power transmission data sequence by using a preset generator to form an estimated value sequence; and generating a cable power transmission data completion result based on the cable power transmission data sequence, the mask feature sequence and the estimated value sequence. Therefore, by implementing the method and the device, the problem of low accuracy of cable completion data in the prior art can be solved.

Description

Technical Field

[0001] This invention relates to the field of cable data technology, and in particular to a method and apparatus for completing cable transmission data. Background Technology

[0002] With the accelerated intelligent and digital transformation of power systems, sensor devices such as phasor measurement units (PMUs) and supervisory control systems (SCADA) deployed in cable transmission networks continuously generate massive amounts of high-dimensional time-series data, providing important data for power system transient stability assessment. However, in actual operation, problems such as sensor failures, communication delays, and electromagnetic interference lead to data loss rates as high as 15%-30%, and the data exhibits strong spatiotemporal randomness and complex loss patterns.

[0003] Traditional data interpolation methods directly fit power grid time series data without considering temporal correlations. This results in overly coarse fitting and significant fluctuations in computational errors. Furthermore, they struggle to effectively model the dynamic coupling relationships within the power grid topology, leading to inaccurate data representation of the actual physical connections and energy flow constraints of the power grid. Therefore, current power grid data completion methods suffer from low accuracy. Summary of the Invention

[0004] This invention provides a method and apparatus for completing cable transmission data, which can solve the problem of low accuracy of cable completion data in the prior art.

[0005] To address the aforementioned technical problems, this invention provides a method for completing cable transmission data, comprising:

[0006] Acquire cable transmission data sequences, power grid topology data, and cable transient timestamps;

[0007] The key cable points in the power grid topology data are compared with the cable locations corresponding to each cable transmission data in the cable transmission data sequence to generate a topology tag sequence for the cable transmission data sequence.

[0008] Based on the cable transient timestamp, determine whether each cable transmission data in the cable transmission data sequence is in a transient process, and generate a transient tag sequence;

[0009] Feature extraction is performed on the cable transmission data sequence to generate a data feature sequence and a mask feature sequence; wherein, the data features are used to indicate whether the cable transmission data is a real value, and the mask features are used to indicate whether the cable transmission data is missing;

[0010] Based on the topology tag sequence, the transient tag sequence, the data feature sequence, and the mask feature sequence, the cable transmission data sequence is subjected to tag matching processing to generate cable analysis data.

[0011] Based on the cable analysis data, a preset generator is used to estimate the cable transmission data sequence to form an estimated value sequence.

[0012] Based on the cable transmission data sequence, the mask feature sequence, and the estimated value sequence, a missing data sequence and an observation data sequence are generated;

[0013] Based on the missing data sequence and the observed data sequence, a cable transmission data completion result is generated.

[0014] As a preferred embodiment, the step of comparing key cable points in the power grid topology data with the cable locations corresponding to each cable transmission data point in the cable transmission data sequence to generate a topology tag sequence for the cable transmission data sequence includes:

[0015] A power grid topology map is constructed based on the power grid topology data;

[0016] The number of cable connections at each power grid location is determined based on the node connection relationships in the power grid topology diagram.

[0017] Locations where the number of cable connections exceeds a preset connection threshold are identified as critical cable points;

[0018] Determine whether each cable transmission data point in the cable transmission data sequence is at a critical cable point, and generate a topology label for each cable transmission data point; wherein, the topology label includes critical cable points and non-critical cable points;

[0019] Based on the topology tags of each of the cable transmission data, a topology tag sequence of the cable transmission data sequence is generated.

[0020] As a preferred embodiment, the step of extracting features from the cable transmission data sequence to generate a data feature sequence and a mask feature sequence includes:

[0021] Using the RW sliding window method, based on the data source of the cable transmission data sequence, it is determined whether each cable transmission data in the cable transmission data sequence is a true value, and a data feature sequence is generated;

[0022] The cable transmission data sequence is traversed based on the data acquisition timestamp to determine whether there are missing data in the cable transmission data sequence, and a mask feature sequence is generated.

[0023] As a preferred embodiment, the step of estimating the cable transmission data sequence based on the cable analysis data using a preset generator to form an estimated value sequence includes:

[0024] Using a preset generator, based on the data feature sequence, the true values ​​are filtered out from the cable transmission data sequence and determined as the estimated dependency data;

[0025] Based on the topology label sequence and the transient label sequence, the cable transmission data sequence is classified into critical transient missing data, critical non-transient missing data, non-critical transient missing data, and non-critical non-transient missing data.

[0026] Based on the estimated dependency data, data estimation is performed on the key transient missing data, the key non-transient missing data, the non-key transient missing data, and the non-key non-transient missing data respectively to generate an estimated value sequence.

[0027] As a preferred embodiment, the step of estimating data based on the estimated dependency data, specifically the critical transient missing data, the critical non-transient missing data, the non-critical transient missing data, and the non-critical non-transient missing data, to generate an estimated value sequence, includes:

[0028] The topological adjacent cable transmission data and time-series adjacent cable transmission data of the key transient missing data are determined from the estimated dependent data, and the key transient missing data are estimated to form the first estimated data;

[0029] The topological adjacent cable transmission data of the key non-transient missing data are determined from the estimated dependent data, and data estimation is performed on the key non-transient missing data to form the second estimated data;

[0030] The time-series adjacent cable transmission data of the non-critical transient missing data are determined from the estimated dependent data, and data estimation is performed on the non-critical transient missing data to form the third estimated data;

[0031] Based on the estimated dependency data, the non-critical and non-transient missing data are estimated to form the fourth estimated data.

[0032] A sequence of estimated values ​​is generated based on the first estimated data, the second estimated data, the third estimated data, and the fourth estimated data.

[0033] As a preferred embodiment, the step of generating the missing data sequence and the observation data sequence based on the cable transmission data sequence, the mask feature sequence, and the estimated value sequence includes:

[0034] The missing data sequence and the observed data sequence are calculated using the following formula:

[0035]

[0036] In the formula, M L For missing data sequences; Mo X is the mask feature sequence; N is the mask feature sequence; X imp X is the estimated value sequence; X is the cable transmission data sequence.

[0037] As a preferred embodiment, generating cable transmission data completion results based on the missing data sequence and the observed data sequence includes:

[0038] Based on the missing data sequence and the observed data sequence, a complete imputation sequence is generated;

[0039] The complete interpolation sequence is determined as the cable transmission data completion result;

[0040] The complete interpolation sequence is calculated using the following formula:

[0041] M = M L +M o

[0042] In the formula, M is the complete interpolation sequence; M L For missing data sequences; M o This is a mask feature sequence.

[0043] As a preferred embodiment, after generating the cable transmission data completion result based on the missing data sequence and the observed data sequence, the method further includes:

[0044] Generate a prompt sequence based on the mask feature sequence and the random binary matrix;

[0045] The preset generator is updated based on the missing data sequence and the prompt sequence.

[0046] As a preferred embodiment, the step of generating the prompt sequence based on the mask feature sequence and the random binary matrix includes:

[0047] The prompt sequence is calculated using the following formula:

[0048] H = B⊙N + 1 / 2(1-B)

[0049] In the formula, H is the prompt sequence; B is a random binary matrix; and N is the mask feature sequence.

[0050] Accordingly, the present invention provides a cable transmission data completion device, comprising: a data acquisition module, a topology analysis module, a transient analysis module, a feature extraction module, an analysis data generation module, a data estimation module, a missing data generation module, and a data completion module;

[0051] The data acquisition module is used to acquire cable transmission data sequences, power grid topology data, and cable transient timestamps.

[0052] The topology analysis module is used to compare the key cable points in the power grid topology data with the cable positions corresponding to each cable transmission data in the cable transmission data sequence, and generate a topology tag sequence for the cable transmission data sequence.

[0053] The transient analysis module is used to determine whether each cable transmission data in the cable transmission data sequence is in a transient process based on the cable transient timestamp, and to generate a transient tag sequence.

[0054] The feature extraction module is used to extract features from the cable transmission data sequence to generate a data feature sequence and a mask feature sequence; wherein, the data features are used to indicate whether the cable transmission data is a real value, and the mask features are used to indicate whether the cable transmission data is missing;

[0055] The analysis data generation module is used to perform tag matching processing on the cable transmission data sequence based on the topology tag sequence, the transient tag sequence, the data feature sequence, and the mask feature sequence to generate cable analysis data.

[0056] The data estimation module is used to estimate the cable transmission data sequence based on the cable analysis data and using a preset generator to form an estimated value sequence.

[0057] The missing data generation module is used to generate a missing data sequence and an observation data sequence based on the cable transmission data sequence, the mask feature sequence, and the estimated value sequence;

[0058] The data completion module is used to generate cable transmission data completion results based on the missing data sequence and the observed data sequence.

[0059] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0060] This invention provides a method for cable transmission data completion, which involves acquiring cable transmission data sequences, power grid topology data, and cable transient timestamps; generating a topology label sequence for the cable transmission data sequences based on the power grid topology data; generating a transient label sequence for the cable transmission data sequences based on the cable transient timestamps; extracting features from the cable transmission data sequences to generate data feature sequences and mask feature sequences; performing label matching processing on the cable transmission data sequences based on the topology label sequence, transient label sequence, data feature sequence, and mask feature sequence to generate cable analysis data; using the cable analysis data, estimating the cable transmission data sequences using a preset generator to form an estimated value sequence; generating a missing data sequence and an observed data sequence based on the cable transmission data sequences, mask feature sequences, and estimated value sequences; and generating cable transmission data completion results based on the missing data sequence and observed data sequences. This invention acquires power grid topology data and cable transient timestamps. When using a generator to estimate data, it comprehensively considers the temporal correlation between the power grid topology and cable transmission data, so that the completed data can adapt to the actual physical connection of the power grid and more closely match the temporal logic of the real data, thereby improving the accuracy of cable completion data. Attached Figure Description

[0061] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0062] Figure 1 A flowchart illustrating one embodiment of the cable transmission data completion method provided by the present invention;

[0063] Figure 2 A schematic diagram of a network architecture for the discriminator provided by the present invention;

[0064] Figure 3 This is a schematic diagram of one embodiment of the cable transmission data completion device provided by the present invention. Detailed Implementation

[0065] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0066] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0067] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0068] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0069] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0070] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0071] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0072] See Figure 1To address the problem of low accuracy in cable transmission data completion in existing technologies, an embodiment of the present invention provides a cable transmission data completion method, which includes steps 101 to 108, the specific details of which are as follows:

[0073] Step 101: Obtain the cable transmission data sequence, power grid topology data, and cable transient timestamps.

[0074] In this embodiment of the invention, the cable transmission data sequence includes cable transmission data at multiple time points for each cable location within the power grid. When missing data exists in the cable transmission data sequence, more suitable data can be selected for data completion by considering the power grid topology of the cable and the transient timestamp of the cable during a transient process. Therefore, by collecting power grid topology data and cable transient timestamps, the accuracy of data completion for the cable transmission data sequence can be improved.

[0075] In this embodiment of the invention, because time-series data possesses high-dimensional characteristics, the redundancy, noise interference, and computational burden of the data also increase significantly. Therefore, before data completion, the cable transmission data sequence is first subjected to dimensionality reduction processing. PCA analysis can be used to reduce the dimensionality of the cable transmission data sequence. The PCA algorithm is based on linear mapping, performing low-dimensional mapping processing on high-dimensional data, and reducing the dimensionality of the data group by maximizing the projection variance. The specific steps are as follows:

[0076] First, the data set is decentralized using the difference between the data means:

[0077]

[0078] In the formula, x is the mean vector of the dataset obtained through decentralization; i Let be the i-th sample in the original dataset; n is the total number of samples.

[0079] Next, the eigenvalues ​​and eigenvectors of the covariance are calculated to complete the data mapping:

[0080]

[0081] In the formula, X i Sum of Y i Let X and Y be the values ​​of the i-th sample on features X and Y, respectively. and Let X and Y be the means of features X and Y.

[0082] Based on the above calculation method, the covariance matrix C is obtained:

[0083]

[0084] The goal of eigenvalue decomposition of the covariance matrix C is to find the eigenvalues ​​λ and eigenvectors v that satisfy:

[0085] Cv i =λ i v i

[0086] The solution to the characteristic equation is:

[0087] det(c-λI)=0

[0088] Standardized data X 标准化 Projected onto principal component directions:

[0089] Z = X 标准化 ·V k

[0090] In the formula, X 标准化 The original data matrix (n×d) is standardized, with each row representing a sample and each column representing a feature. V k Z is the projection matrix (d×k) composed of the first k eigenvectors (principal component directions). Z is the dimensionality-reduced data matrix (n×k), where each sample is compressed to k principal component dimensions.

[0091] The feature vectors are standardized to reduce the data dimension n to k, and the feature dimensions with zero variance are eliminated to achieve data feature dimensionality reduction.

[0092]

[0093] In the formula, x is the gray value of the original pixel; m is the average gray value of the pixel; s is the variance of the pixel gray value; and N is the standardized new pixel value.

[0094] Step 102: Compare the key cable points in the power grid topology data with the cable locations corresponding to each cable transmission data in the cable transmission data sequence to generate a topology tag sequence for the cable transmission data sequence.

[0095] As a preferred embodiment, the step of comparing the key cable points in the power grid topology data with the cable locations corresponding to each cable transmission data in the cable transmission data sequence to generate a topology tag sequence for the cable transmission data sequence includes:

[0096] A power grid topology map is constructed based on the power grid topology data;

[0097] The number of cable connections at each power grid location is determined based on the node connection relationships in the power grid topology diagram.

[0098] Locations where the number of cable connections exceeds a preset connection threshold are identified as critical cable points;

[0099] Determine whether each cable transmission data point in the cable transmission data sequence is at a critical cable point, and generate a topology label for each cable transmission data point; wherein, the topology label includes critical cable points and non-critical cable points;

[0100] Based on the topology tags of each of the cable transmission data, a topology tag sequence of the cable transmission data sequence is generated.

[0101] In this embodiment of the invention, multiple key cable points can be identified by analyzing the topology data of the power grid where the cable is located. Specifically, a power grid topology map is constructed based on the power grid topology data. Nodes (such as substations and switchgear) and their connections (such as cable lines) are extracted from the cable topology data to construct the power grid topology map (nodes are vertices, and cables are edges). The number of cable connections at each node in the power grid topology map is analyzed. Since a node with more cable connections means it has more direct electrical connections with other nodes, and key cable points often serve as power transmission hubs, connecting multiple different transmission areas, nodes with more than a preset connection threshold can be identified as key cable points. Cable transmission data is collected from the locations of various cables in the power grid. Therefore, by comparing the key cable points with the corresponding cable locations in the cable transmission data, topology tags for each cable transmission data point can be obtained. Collecting the topology tags of each cable transmission data point generates a topology tag sequence for the cable transmission data sequence.

[0102] Step 103: Based on the cable transient timestamp, determine whether each cable transmission data in the cable transmission data sequence is in a transient process, and generate a transient tag sequence.

[0103] In this embodiment of the invention, each cable transmission data corresponds to a data acquisition time, and the cable transient timestamp records multiple time points when the cable is in a transient state. Therefore, by comparing the cable transient timestamp with the data acquisition time corresponding to each cable transmission data, the transient label of each cable transmission data can be obtained. The transient label includes transient and non-transient states. The transient labels of each cable transmission data can be combined to generate a transient label sequence of the cable transmission data sequence.

[0104] Step 104: Extract features from the cable transmission data sequence to generate a data feature sequence and a mask feature sequence; wherein, the data features are used to indicate whether the cable transmission data is a true value, and the mask features are used to indicate whether the cable transmission data is missing.

[0105] As a preferred embodiment, feature extraction is performed on the cable transmission data sequence to generate a data feature sequence and a mask feature sequence, including:

[0106] Using the RW sliding window method, based on the data source of the cable transmission data sequence, it is determined whether each cable transmission data in the cable transmission data sequence is a true value, and a data feature sequence is generated;

[0107] The cable transmission data sequence is traversed based on the data acquisition timestamp to determine whether there are missing data in the cable transmission data sequence, and a mask feature sequence is generated.

[0108] In this embodiment of the invention, the RW sliding window method is used. By sliding a window of size r, data feature sequences and mask feature sequences can be extracted through dual channels of data features and mask features. Data feature I1 and mask feature I2 explicitly include the spatiotemporal distribution information of missing positions, improving the learning rate of time-series information features. The mask features within the feature map contain missing information to address the problem of poor data temporal uniformity. Data feature I1 indicates whether the cable transmission data is a true value; if the data is an original, collected true value, it is marked as 1; if it is a temporary interpolation value, it is marked as 0. These 0 / 1 marks are arranged in chronological order to form a data feature sequence. The source of the cable transmission data sequence can determine whether the cable transmission data is a true value or a temporary interpolation value. Mask feature I2 indicates whether the cable transmission data is missing. By traversing the data within the window, if the data is missing (not collected or invalid), it is marked as 0; if it is complete and valid, it is marked as 1. These 0 / 1 marks are arranged in chronological order to form a mask feature sequence.

[0109] Step 105: Based on the topology tag sequence, the transient tag sequence, the data feature sequence, and the mask feature sequence, perform tag matching processing on the cable transmission data sequence to generate cable analysis data.

[0110] In this embodiment of the invention, the topology tag sequence, transient tag sequence, data feature sequence, and mask feature sequence are all sequences arranged in chronological order. Similarly, the cable transmission data sequence is also formed by arranging the cable transmission data in chronological order. Therefore, by performing tag matching between the topology tag sequence, transient tag sequence, data feature sequence, and mask feature sequence and the cable transmission data sequence, corresponding topology tags, transient tags, data features, and mask features can be matched to each cable transmission data point. Arranging the cable transmission data with matched tags and features in chronological order generates cable analysis data.

[0111] Step 106: Based on the cable analysis data, use a preset generator to estimate the cable transmission data sequence to form an estimated value sequence.

[0112] In this embodiment of the invention, cable analysis data is input into a preset generator, which learns the potential distribution of complete power grid time-series data and outputs a complete estimated value sequence with the same dimension as the original cable transmission data. The data feature sequence provides the generator with a priori data distribution guided by physical laws, while the mask feature sequence guides the generator to focus on the spatiotemporal location to be interpolated.

[0113] As a preferred embodiment, based on the cable analysis data, a preset generator is used to estimate the cable transmission data sequence to form an estimated value sequence, including:

[0114] Using a preset generator, based on the data feature sequence, the true values ​​are filtered out from the cable transmission data sequence and determined as the estimated dependency data;

[0115] Based on the topology label sequence and the transient label sequence, the cable transmission data sequence is classified into critical transient missing data, critical non-transient missing data, non-critical transient missing data, and non-critical non-transient missing data.

[0116] Based on the estimated dependency data, data estimation is performed on the key transient missing data, the key non-transient missing data, the non-key transient missing data, and the non-key non-transient missing data respectively to generate an estimated value sequence.

[0117] In this embodiment of the invention, based on the data feature sequence, it can be determined which values ​​in the cable transmission data sequence are true values ​​and which are temporary interpolated values. Therefore, the true values ​​in the cable transmission data sequence are filtered out and can be used as the data basis for data estimation, forming estimation dependency data. Based on the topology label sequence and transient label sequence, data at critical cable points and data in transient processes in the cable transmission data sequence can be identified. Therefore, the sequences in the cable transmission data sequence can be classified into: critical transient missing data that is both at critical cable points and in transient processes; critical non-transient missing data that is at critical cable points but not in transient processes; non-critical transient missing data that is not at critical cable points but in transient processes; and non-critical non-transient missing data that is neither at critical cable points nor in transient processes. Based on the characteristics of these four types of data, appropriate data is selected from the estimation dependency data for data estimation, and estimated values ​​for each of the four types of data can be obtained, thus forming an estimation value sequence.

[0118] As a preferred embodiment, based on the estimated dependency data, data estimation is performed on the critical transient missing data, the critical non-transient missing data, the non-critical transient missing data, and the non-critical non-transient missing data respectively to generate an estimated value sequence, including:

[0119] The topological adjacent cable transmission data and time-series adjacent cable transmission data of the key transient missing data are determined from the estimated dependent data, and the key transient missing data are estimated to form the first estimated data;

[0120] The topological adjacent cable transmission data of the key non-transient missing data are determined from the estimated dependent data, and data estimation is performed on the key non-transient missing data to form the second estimated data;

[0121] The time-series adjacent cable transmission data of the non-critical transient missing data are determined from the estimated dependent data, and data estimation is performed on the non-critical transient missing data to form the third estimated data;

[0122] Based on the estimated dependency data, the non-critical and non-transient missing data are estimated to form the fourth estimated data.

[0123] A sequence of estimated values ​​is generated based on the first estimated data, the second estimated data, the third estimated data, and the fourth estimated data.

[0124] In a preferred embodiment, for cable transmission data located at critical cable points, data estimation is performed by strengthening the data dependencies between adjacent nodes. Therefore, data estimation is performed based on the cable transmission data adjacent to the topology of this cable transmission data, thus obtaining the corresponding estimated value. For cable transmission data in a transient process, data estimation is performed by strengthening the data dependencies within the temporal context. Therefore, data estimation is performed based on the cable transmission data adjacent to the topology of this cable transmission data, thus obtaining the corresponding estimated value. For non-critical, non-transient missing data that is neither located at critical cable points nor in a transient process, a preset generator is used to perform data estimation based on the estimated dependency data. After obtaining the estimated values ​​of each data point in the cable transmission data sequence, they are arranged in chronological order to form an estimated value sequence. The generation of the estimated value sequence can be expressed by the following formula:

[0125] X imp =G([I1, I2], θ) G )

[0126] In the formula, X imp The estimated value sequence consists of the generator's estimates of data at all positions within the entire time window (regardless of whether the original data is missing); I1 is the data feature sequence; I2 is the mask feature sequence; θ G G represents the network parameters of the generator; G(·) is the data estimation function.

[0127] Step 107: Based on the cable transmission data sequence, the mask feature sequence, and the estimated value sequence, generate the missing data sequence and the observation data sequence.

[0128] As a preferred embodiment, based on the cable transmission data sequence, the mask feature sequence, and the estimated value sequence, a missing data sequence and an observation data sequence are generated, including:

[0129] The missing data sequence and the observed data sequence are calculated using the following formula:

[0130]

[0131] In the formula, M L For missing data sequences; M o X is the mask feature sequence; N is the mask feature sequence; X imp X is the estimated value sequence; X is the cable transmission data sequence.

[0132] In this embodiment of the invention, the complete interpolation sequence is composed of a missing data sequence and an observed data sequence. The missing data sequence is formed by filling in the missing locations based on the estimated values ​​of the missing locations output by the generator. The observed data sequence retains the data that is not missing from the original cable transmission data; the data at these locations is accurate and reliable, and does not require generator completion.

[0133] Step 108: Based on the missing data sequence and the observed data sequence, generate cable transmission data completion results.

[0134] As a preferred embodiment, based on the missing data sequence and the observed data sequence, a cable transmission data completion result is generated, including:

[0135] Based on the missing data sequence and the observed data sequence, a complete imputation sequence is generated;

[0136] The complete interpolation sequence is determined as the cable transmission data completion result;

[0137] The complete interpolation sequence is calculated using the following formula:

[0138] M = M L +M o

[0139] In the formula, M is the complete interpolation sequence; M L For missing data sequences; M o This is a mask feature sequence.

[0140] In this embodiment of the invention, the cable transmission data completion result consists of the original cable transmission data that is not missing and the missing data filled in by the generator's estimated values. Therefore, by adding the missing data sequence and the observed data sequence, a complete interpolation sequence can be formed, thereby obtaining the cable transmission data completion result.

[0141] As a preferred embodiment, after generating the cable transmission data completion result based on the missing data sequence and the observed data sequence, the method further includes:

[0142] Generate a prompt sequence based on the mask feature sequence and the random binary matrix;

[0143] The preset generator is updated based on the missing data sequence and the prompt sequence.

[0144] As a preferred embodiment, generating a prompt sequence based on the mask feature sequence and the random binary matrix includes:

[0145] The prompt sequence is calculated using the following formula:

[0146] H = B⊙N + 1 / 2(1-B)

[0147] In the formula, H is the prompt sequence; B is a random binary matrix; and N is the mask feature sequence.

[0148] In this embodiment of the invention, a GAIN-W discriminator is constructed using three sets of fully connected layers, which can solve the gradient explosion problem. Then, the generator is guided to learn based on the missing data sequence and the prompt data to improve the data estimation accuracy of the generator.

[0149] See Figure 2 This is a schematic diagram of a network architecture for the discriminator provided by the present invention. The input data of the discriminator is interpolated data M. L And cue data H, the introduction of cue data H is used to guide the generator to learn the data distribution. M L The discriminator's output is composed of the original missing parts and the complement values ​​corresponding to the missing positions in the estimated value sequence output by the generator. B is a random binary (0,1) matrix; when B=1, the corresponding position requires a cue data H for information. The discriminator's input data undergoes dimensionality reduction, feature extraction, and linear transformation through a fully connected layer. The ReLU activation function performs nonlinear correction on the linear transformation result output by the fully connected layer, distinguishing between the actual observed data and the generated imputed data, thus generating the discriminator's coupled output C. The discriminator uses the Wasserstein value to reflect the loss between the learned distribution and the true distribution, used to distinguish between imputed data and observed data. The loss function is as follows:

[0150] L D =(1-N)⊙CN⊙C

[0151] In the formula, L D denoted as the discriminator loss; N is the mask feature sequence; and C is the discriminator coupling output.

[0152] Implementing the above embodiments has the following effects:

[0153] This invention provides a method for cable transmission data completion, which involves acquiring cable transmission data sequences, power grid topology data, and cable transient timestamps; generating a topology label sequence for the cable transmission data sequences based on the power grid topology data; generating a transient label sequence for the cable transmission data sequences based on the cable transient timestamps; extracting features from the cable transmission data sequences to generate data feature sequences and mask feature sequences; performing label matching processing on the cable transmission data sequences based on the topology label sequence, transient label sequence, data feature sequence, and mask feature sequence to generate cable analysis data; using the cable analysis data, estimating the cable transmission data sequences using a preset generator to form an estimated value sequence; generating a missing data sequence and an observed data sequence based on the cable transmission data sequences, mask feature sequences, and estimated value sequences; and generating cable transmission data completion results based on the missing data sequence and observed data sequences. This invention acquires power grid topology data and cable transient timestamps. When using a generator to estimate data, it comprehensively considers the temporal correlation between the power grid topology and cable transmission data, so that the completed data can adapt to the actual physical connection of the power grid and more closely match the temporal logic of the real data, thereby improving the accuracy of cable completion data.

[0154] like Figure 3 As shown, based on the above method embodiments, corresponding apparatus embodiments are provided;

[0155] One embodiment of the present invention provides a cable transmission data completion device, comprising: a data acquisition module, a topology analysis module, a transient analysis module, a feature extraction module, an analysis data generation module, a data estimation module, a missing data generation module, and a data completion module;

[0156] The data acquisition module is used to acquire cable transmission data sequences, power grid topology data, and cable transient timestamps.

[0157] The topology analysis module is used to compare the key cable points in the power grid topology data with the cable positions corresponding to each cable transmission data in the cable transmission data sequence, and generate a topology tag sequence for the cable transmission data sequence.

[0158] The transient analysis module is used to determine whether each cable transmission data in the cable transmission data sequence is in a transient process based on the cable transient timestamp, and to generate a transient tag sequence.

[0159] The feature extraction module is used to extract features from the cable transmission data sequence to generate a data feature sequence and a mask feature sequence; wherein, the data features are used to indicate whether the cable transmission data is a real value, and the mask features are used to indicate whether the cable transmission data is missing;

[0160] The analysis data generation module is used to perform tag matching processing on the cable transmission data sequence based on the topology tag sequence, the transient tag sequence, the data feature sequence, and the mask feature sequence to generate cable analysis data.

[0161] The data estimation module is used to estimate the cable transmission data sequence based on the cable analysis data and using a preset generator to form an estimated value sequence.

[0162] The missing data generation module is used to generate a missing data sequence and an observation data sequence based on the cable transmission data sequence, the mask feature sequence, and the estimated value sequence;

[0163] The data completion module is used to generate cable transmission data completion results based on the missing data sequence and the observed data sequence.

[0164] It is understood that the above-described device embodiments correspond to the method embodiments of the present invention, and can implement the cable transmission data completion method provided by any of the above-described method embodiments of the present invention.

[0165] It should be noted that the device embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can specifically be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0166] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A method for completing cable transmission data, characterized in that, include: Acquire cable transmission data sequences, power grid topology data, and cable transient timestamps; The key cable points in the power grid topology data are compared with the cable locations corresponding to each cable transmission data in the cable transmission data sequence to generate a topology tag sequence for the cable transmission data sequence. Based on the cable transient timestamp, determine whether each cable transmission data in the cable transmission data sequence is in a transient process, and generate a transient tag sequence; Feature extraction is performed on the cable transmission data sequence to generate a data feature sequence and a mask feature sequence; wherein, the data features are used to indicate whether the cable transmission data is a real value, and the mask features are used to indicate whether the cable transmission data is missing; Based on the topology tag sequence, the transient tag sequence, the data feature sequence, and the mask feature sequence, the cable transmission data sequence is subjected to tag matching processing to generate cable analysis data. Based on the cable analysis data, a preset generator is used to estimate the cable transmission data sequence to form an estimated value sequence. Based on the cable transmission data sequence, the mask feature sequence, and the estimated value sequence, a missing data sequence and an observation data sequence are generated; Based on the missing data sequence and the observed data sequence, a cable transmission data completion result is generated.

2. The cable transmission data completion method according to claim 1, characterized in that, The step of comparing key cable points in the power grid topology data with the cable locations corresponding to each cable transmission data point in the cable transmission data sequence to generate a topology tag sequence for the cable transmission data sequence includes: A power grid topology map is constructed based on the power grid topology data; The number of cable connections at each power grid location is determined based on the node connection relationships in the power grid topology diagram. Locations where the number of cable connections exceeds a preset connection threshold are identified as critical cable points; Determine whether each cable transmission data point in the cable transmission data sequence is at a critical cable point, and generate a topology label for each cable transmission data point; wherein, the topology label includes critical cable points and non-critical cable points; Based on the topology tags of each of the cable transmission data, a topology tag sequence of the cable transmission data sequence is generated.

3. The cable transmission data completion method according to claim 1, characterized in that, The step of extracting features from the cable transmission data sequence to generate a data feature sequence and a mask feature sequence includes: Using the RW sliding window method, based on the data source of the cable transmission data sequence, it is determined whether each cable transmission data in the cable transmission data sequence is a true value, and a data feature sequence is generated; The cable transmission data sequence is traversed based on the data acquisition timestamp to determine whether there are missing data in the cable transmission data sequence, and a mask feature sequence is generated.

4. The cable transmission data completion method according to claim 1, characterized in that, The step of estimating the cable transmission data sequence based on the cable analysis data using a preset generator to form an estimated value sequence includes: Using a preset generator, based on the data feature sequence, the true values ​​are filtered out from the cable transmission data sequence and determined as the estimated dependency data; Based on the topology label sequence and the transient label sequence, the cable transmission data sequence is classified into critical transient missing data, critical non-transient missing data, non-critical transient missing data, and non-critical non-transient missing data. Based on the estimated dependency data, data estimation is performed on the key transient missing data, the key non-transient missing data, the non-key transient missing data, and the non-key non-transient missing data respectively to generate an estimated value sequence.

5. The cable transmission data completion method according to claim 4, characterized in that, Based on the estimated dependency data, data estimation is performed on the key transient missing data, the key non-transient missing data, the non-key transient missing data, and the non-key non-transient missing data respectively to generate an estimated value sequence, including: The topological adjacent cable transmission data and time-series adjacent cable transmission data of the key transient missing data are determined from the estimated dependent data, and the key transient missing data are estimated to form the first estimated data; The topological adjacent cable transmission data of the key non-transient missing data are determined from the estimated dependent data, and data estimation is performed on the key non-transient missing data to form the second estimated data; The time-series adjacent cable transmission data of the non-critical transient missing data are determined from the estimated dependent data, and data estimation is performed on the non-critical transient missing data to form the third estimated data; Based on the estimated dependency data, the non-critical and non-transient missing data are estimated to form the fourth estimated data. A sequence of estimated values ​​is generated based on the first estimated data, the second estimated data, the third estimated data, and the fourth estimated data.

6. The cable transmission data completion method according to claim 1, characterized in that, The process of generating missing data sequences and observation data sequences based on the cable transmission data sequence, the mask feature sequence, and the estimated value sequence includes: The missing data sequence and the observed data sequence are calculated using the following formula: In the formula, M L For missing data sequences; M o X is the mask feature sequence; N is the mask feature sequence; X imp X is the estimated value sequence; X is the cable transmission data sequence.

7. The cable transmission data completion method according to claim 1, characterized in that, The process of generating cable transmission data completion results based on the missing data sequence and the observed data sequence includes: Based on the missing data sequence and the observed data sequence, a complete imputation sequence is generated; The complete interpolation sequence is determined as the cable transmission data completion result; The complete interpolation sequence is calculated using the following formula: M=M L +M o In the formula, M is the complete interpolation sequence; M L For missing data sequences; M o This is a mask feature sequence.

8. The cable transmission data completion method according to claim 1, characterized in that, After generating the cable transmission data completion result based on the missing data sequence and the observed data sequence, the process further includes: Generate a prompt sequence based on the mask feature sequence and the random binary matrix; The preset generator is updated based on the missing data sequence and the prompt sequence.

9. The cable transmission data completion method according to claim 8, characterized in that, The step of generating a prompt sequence based on the mask feature sequence and the random binary matrix includes: The prompt sequence is calculated using the following formula: H = B☉N + 1 / 2(1-B) In the formula, H is the prompt sequence; B is a random binary matrix; and N is the mask feature sequence.

10. A cable transmission data completion device, characterized in that, include: The module includes: data acquisition module, topology analysis module, transient analysis module, feature extraction module, analysis data generation module, data estimation module, missing data generation module, and data completion module. The data acquisition module is used to acquire cable transmission data sequences, power grid topology data, and cable transient timestamps. The topology analysis module is used to compare the key cable points in the power grid topology data with the cable positions corresponding to each cable transmission data in the cable transmission data sequence, and generate a topology tag sequence for the cable transmission data sequence. The transient analysis module is used to determine whether each cable transmission data in the cable transmission data sequence is in a transient process based on the cable transient timestamp, and to generate a transient tag sequence. The feature extraction module is used to extract features from the cable transmission data sequence to generate a data feature sequence and a mask feature sequence; wherein, the data features are used to indicate whether the cable transmission data is a real value, and the mask features are used to indicate whether the cable transmission data is missing; The analysis data generation module is used to perform tag matching processing on the cable transmission data sequence based on the topology tag sequence, the transient tag sequence, the data feature sequence, and the mask feature sequence to generate cable analysis data. The data estimation module is used to estimate the cable transmission data sequence based on the cable analysis data and using a preset generator to form an estimated value sequence. The missing data generation module is used to generate a missing data sequence and an observation data sequence based on the cable transmission data sequence, the mask feature sequence, and the estimated value sequence; The data completion module is used to generate cable transmission data completion results based on the missing data sequence and the observed data sequence.

Images