A method, apparatus, device and medium for processing power grid equipment data

Abstract

Embodiments of the present application disclose a kind of power grid equipment data processing method, device, equipment and medium.The method comprises: obtaining the observation corresponding to the power grid equipment state data based on measurement matrix, measurement matrix is recovery error optimization measurement matrix;Observation is handled based on subspace tracking algorithm, and the target reconstruction data corresponding to the power grid equipment state data is reconstructed, data is obtained by adaptive measurement matrix, and signal reconstruction is realized by improved subspace tracking algorithm, the existing compressed sensing measurement matrix is random in observation time, except that the problem of over-redundancy of data amount, waste hardware resources cannot be effectively collected to the signal after sparsification, it is realized that in the reservation equipment characteristic information ensures that the reconstruction error of data is low, the data acquisition amount is reduced, effectively reduce the communication overhead, reduce the data processing pressure of data center.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and in particular to a power grid equipment data processing method and device, equipment and medium. BACKGROUND

[0002] Under the background of power internet of things, it is necessary to collect the state parameters and physical characteristics of key nodes such as ring network cabinets, distribution transformers, branch boxes, and electric energy meters at any time, evaluate their operating states, and arrange maintenance and maintenance plans based thereon. However, in order to better obtain the physical characteristics of the operating state of the equipment, high-frequency collection has become mainstream. However, high-frequency collection not only produces a large amount of redundant data, but also produces a large amount of data in a short time, bringing great challenges to data storage and rapid processing. On the other hand, real-time is an important feature of power internet of things state perception, and too much redundant data is not conducive to the extraction of effective features and the mining of useful information, and is not conducive to online monitoring and fault diagnosis of equipment, so how to greatly reduce the data transmission under the premise of ensuring the integrity of the data features is a hot issue to be solved.

[0003] Compressed sensing is a new signal processing technology, which points out that if a signal is sparse or sparse after transformation, it can be projected onto a low dimension by constructing a measurement matrix, realizing the compression of data. Finally, the original signal is reconstructed by observation value using reconstruction algorithm, and then used for data analysis. By adjusting the sampling rate, the amount of data is greatly reduced, and the communication pressure is reduced. Using compressed sensing technology to process the state model of power equipment, while accurately preserving the characteristics of the equipment, the rapid identification of equipment failure is realized.

[0004] The construction of the measurement matrix is a key step of compressed sensing. The existing compressed sensing measurement matrix includes Gaussian measurement matrix, sparse random matrix, and chaotic measurement matrix. However, these measurement matrices are random in observation time, and in addition to being unable to effectively collect the sparse signal, there is also a situation of excessive redundancy of data, wasting hardware resources. In the traditional method, the sampling matrix is generated offline, and a random matrix is usually preferred because such a matrix has an overwhelming probability of limiting equidistance characteristics, which ensures the stability of the recovery algorithm. However, this pre-defined random matrix may waste valuable sampling resources. SUMMARY

[0005] The present application provides a power grid equipment data processing method, device, equipment and medium to realize data compression and improve data reconstruction accuracy.

[0006] According to an aspect of the present application, a power grid equipment data processing method is provided, comprising:

[0007] The observation value determination module is configured to acquire observation values corresponding to the power grid equipment state data based on a measurement matrix, wherein the measurement matrix is a recovery error optimized measurement matrix.

[0008] The data reconstruction module is configured to reconstruct target reconstruction data corresponding to the power grid equipment state data based on a subspace pursuit algorithm.

[0009] According to another aspect of the present application, there is provided a power grid equipment data processing apparatus, comprising:

[0010] The observation value determination module is configured to acquire observation values corresponding to the power grid equipment state data based on a measurement matrix, wherein the measurement matrix is a recovery error optimized measurement matrix.

[0011] The data reconstruction module is configured to reconstruct target reconstruction data corresponding to the power grid equipment state data based on a subspace pursuit algorithm.

[0012] According to another aspect of the present application, there is provided an electronic device, comprising:

[0013] at least one processor; and

[0014] a memory in communication with the at least one processor; wherein

[0015] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the power grid equipment data processing method according to any one of the embodiments of the present application.

[0016] According to another aspect of the present application, there is provided a computer readable storage medium storing computer instructions for enabling a processor to execute the power grid equipment data processing method according to any one of the embodiments of the present application.

[0017] The technical solution of the embodiments of the present application acquires observation values corresponding to the power grid equipment state data based on a measurement matrix, wherein the measurement matrix is a recovery error optimized measurement matrix, and reconstructs target reconstruction data corresponding to the power grid equipment state data based on a subspace pursuit algorithm, thereby solving the problem that the existing compressed sensing measurement matrix is random in observation time, which not only cannot effectively collect the sparse signal, but also has excessive redundant data and wastes hardware resources, and achieving the low reconstruction error of the device feature information, the low data collection amount, the effective reduction of the communication overhead, the reduction of the data processing pressure of the data center, and the improvement of the state monitoring and fault diagnosis capability of the multi-source heterogeneous equipment of the low-voltage power distribution Internet of Things.

[0018] It is to be understood that the description of the background of the application is not an acknowledgement or admission that any of the information provided in the description of the background of the application is prior art to the application. The information in the description of the background of the application may contain ideas, concepts and / or discoveries not yet known to be prior art to the present application. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0020] Figure 1 A flow chart of a power grid equipment data processing method provided by the first embodiment of the present application;

[0021] Figure 2 A process chart of a compressed sensing signal observation based on an adaptive measurement matrix provided by the second embodiment of the present application;

[0022] Figure 3 An improved SP reconstruction algorithm flow provided by the second embodiment of the present application;

[0023] Figure 4 A reconstruction flow chart provided by the second embodiment of the present application;

[0024] Figure 5 A measurement error chart of different measurement matrix construction algorithms provided by the second embodiment of the present application;

[0025] Figure 6 An anti-noise performance chart of actual deployment provided by the second embodiment of the present application;

[0026] Figure 7 A structural schematic diagram of a power grid equipment data processing device provided by the third embodiment of the present application;

[0027] Figure 8 A structural schematic diagram of an electronic device provided by the fourth embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to make the technical personnel in the art better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should be within the scope of protection of the present application.

[0029] It should be noted that the terms "first", "second", and the like in the description and claims of the application and the above drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the application described herein can be implemented in other than the order illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a list of steps or units does not necessarily limit to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to such processes, methods, products or devices.

[0030] Before introducing the technical solution, the application scenario is exemplarily described. In the background of power internet of things, the state parameters and physical characteristics of key nodes such as ring network cabinet, distribution transformer, branch box and electric energy meter need to be collected at any time, the running state is evaluated, and the maintenance and maintenance plan is arranged on this basis. In order to better obtain the physical characteristics of the running state of the equipment, high-frequency collection has become the mainstream. However, high-frequency collection not only produces a large amount of redundant data, but also produces a large amount of data in a short time, which brings great challenges to data storage and rapid processing. On the other hand, real-time is an important feature of power internet of things state perception, too much redundant data is not conducive to the extraction of effective features and the mining of useful information, and is not conducive to the online monitoring and fault diagnosis of equipment. Therefore, the scheme of the embodiment of the application, through the compression processing of the data of the power equipment by the compression sensing technology, and based on the reconstructed algorithm and the compressed data, the original data of the power equipment is reconstructed, which can greatly reduce the transmission amount of data on the premise of ensuring the integrity of the data, and provide important data support for the subsequent fault identification and online monitoring of the power equipment.

[0031] Embodiment one

[0032] Figure 1 The flowchart of the method for processing power grid equipment data provided by the embodiment one of the application, the embodiment can be applicable to the reconstruction of multi-source power distribution big data, the method can be executed by a power grid equipment data processing device, the power grid equipment data processing device can be realized in the form of hardware and / or software, and the power grid equipment data processing device can be configured in a power equipment or a computer equipment. As shown in the figure, the method comprises: Figure 1

[0033] S110, obtaining observation values corresponding to the state data of the power grid equipment based on the measurement matrix, the measurement matrix being a recovery error optimization measurement matrix.

[0034] ​The measurement matrix is an important part of compressed sensing, and is used to realize the collection and compression of power grid equipment state data. The existing measurement matrix includes a Gaussian measurement matrix, a sparse random matrix, a chaotic measurement matrix, and the like. The measurement matrix in the embodiment refers to a recovery error optimization measurement matrix. The power grid equipment state data can be data related to the operating state of power equipment in the power grid, such as voltage and current. The observation value can be understood as data obtained by compressing the power grid equipment state data based on the measurement matrix. The recovery error optimization measurement matrix refers to that the measurement matrix can be optimized by minimizing the recovery error, that is, the measurement matrix in the embodiment can be continuously optimized.

[0035] It should be noted that compressed sensing is a brand-new signal processing technology. It is pointed out that if a signal is sparse or sparse after transformation, the signal can be projected onto a low dimension by constructing a measurement matrix, so as to realize the compression of data. Finally, the original signal is reconstructed by using a reconstruction algorithm through the observation value, and then used for data analysis. By adjusting the sampling rate, the amount of data is greatly reduced by using the compressed sensing technology, and the communication pressure is reduced. The state of the power equipment is processed by using the compressed sensing technology, so as to accurately retain the characteristics of the equipment and realize the rapid identification of the equipment fault.

[0036] It can be understood that the existing measurement matrix includes a Gaussian measurement matrix, a sparse random matrix, a chaotic measurement matrix, and the like. However, these measurement matrices are random in observation time, and in addition to being unable to effectively collect the sparse signal, there is also a problem of excessive redundancy of data, which wastes hardware resources. In the traditional method, the sampling matrix is generated offline, and a random matrix is usually preferred because such a matrix has an overwhelming probability of limiting equidistance characteristics, which guarantees the stability of the recovery algorithm. However, such a predefined random matrix can waste valuable sampling resources. Intuitively, when some initial measurement values are obtained and the basic information of the sparse signal is mastered, the focus can be placed on the entries that are likely to be non-zero, in order to improve the sensing efficiency and the signal-to-noise ratio of each entry. There is no need to allocate sampling work to zero entries. Such a sensing strategy is called adaptive compressed sensing, in which the sensing matrix is designed online and depends on the previous recovery result. Adaptive compressed sensing has shown the potential to improve the recovery accuracy. In the embodiment, a new compressed sensing adaptive measurement matrix generation method is proposed, which optimizes the sampling matrix by minimizing the recovery error. Compared with the existing measurement matrix generation algorithm, the measurement matrix constructed by the algorithm can have a lower recovery error.

[0037] On the basis, the measurement matrix is used to obtain observation values corresponding to the power grid equipment state data, and the measurement matrix is a recovery error optimization measurement matrix, comprising: determining a mean square error function based on voltage data, current data of the power grid equipment and a Cramer-Rao bound criterion; determining the mean square error function as a target function, and generating the measurement matrix based on the target function and a Gaussian matrix.

[0038] The mean square error function refers to the mean square error of the minimum recovery result of the voltage and current data.

[0039] Specifically, the mean square error function is determined based on the Cramer-Rao bound criterion and the voltage and current data of the power grid equipment, and is used as a target function. Then, an adaptive observation matrix is generated through the target function. The adaptive constructed perception matrix helps to improve the performance of the SP algorithm. The benefits come from two aspects: enhanced signal-to-noise ratio and reduced interference between original signals x.

[0040] S120, processing the observation values based on a subspace pursuit algorithm to obtain target reconstruction data corresponding to the power grid equipment state data.

[0041] The subspace algorithm in this embodiment refers to an improved subspace algorithm, which is used to realize the reconstruction of the power grid equipment state data. The target reconstruction data refers to the data corresponding to the power grid equipment state data obtained by reconstruction.

[0042] Specifically, the power grid equipment state information is observed through the measurement matrix to obtain observation values corresponding to the power grid equipment state data. Further, the observation values are processed through the subspace algorithm to realize the reconstruction of the power grid equipment state data. The data obtained by reconstruction is the target reconstruction data.

[0043] It should be further pointed out that the subspace pursuit algorithm in this embodiment can be an improved SP algorithm. The existing reconstruction algorithm is mainly. However, the direct reconstruction of the collected data using the Orthoganal Matching Pursuit (OMP) algorithm and the Subspace Pursuit (SP) algorithm does not consider the correlation between the data, resulting in low reconstruction accuracy and efficiency, which is difficult to meet the real-time monitoring demand of massive data of medium and low voltage distribution networks. The improved SP algorithm in this embodiment uses the prior information estimated in advance to more accurately reconstruct the sampling data at the next moment.

[0044] On the basis, after the target reconstruction data corresponding to the power grid equipment state data is obtained by reconstruction, it further comprises: determining the operation data of the power grid equipment based on the target reconstruction data; if the operation data is consistent with the preset fault data, it is determined that the power grid equipment is in a fault state.

[0045] It can be understood that after obtaining the target reconstruction data, the running state of each power device in the power grid can be determined according to the reconstruction data, such as real-time voltage, current, active power and the like during running. Based on the above running data, by comparing with the preset fault data, it can be determined whether the power device is in a fault state or a non-fault state. For example, when the voltage of the power device is out of limit, it indicates that the line related to the device has a problem, and timely maintenance and repair of the power device are realized.

[0046] On the basis, the target reconstruction data corresponding to the power grid device state data is reconstructed by processing the observation value based on the subspace pursuit algorithm, comprising: predicting the sparsity of the power grid device state data based on the subspace pursuit algorithm; obtaining the to-be-used reconstruction data by processing the observation value based on the sparsity, and determining the reconstruction error of the to-be-used reconstruction data and the power grid device state data.

[0047] Among them, the to-be-used reconstruction data refers to the data corresponding to the power grid device state data reconstructed based on the observation value, and the precision of the to-be-used reconstruction data may not meet the standard. Based on this, the sparsity of the power grid device state data is estimated by the improved subspace pursuit algorithm, and then the power grid device state data is reconstructed in an iterative form to obtain the to-be-used reconstruction data. In order to determine whether the precision of the to-be-used reconstruction data is qualified, the reconstruction error of the to-be-used reconstruction data and the power grid device state data can be calculated.

[0048] On the basis, the target reconstruction data corresponding to the power grid device state data is reconstructed by processing the observation value based on the subspace pursuit algorithm, comprising: when the reconstruction error of the to-be-used reconstruction data is less than a preset reconstruction error threshold, the to-be-used reconstruction data is determined as the target reconstruction data.

[0049] Among them, the preset reconstruction error threshold refers to a threshold set in advance, which can be set by the research and development personnel based on experience.

[0050] Specifically, if the reconstruction error is less than the preset reconstruction error threshold, it indicates that the reconstruction precision meets the requirements, and the to-be-used reconstruction data at this time can be used as the target reconstruction data.

[0051] On the basis, the observation value is processed based on the subspace pursuit algorithm, target reconstruction data corresponding to the power grid equipment state data is obtained through reconstruction, and the target reconstruction data corresponding to the to-be-used reconstruction data is determined through the particle swarm optimization algorithm if the reconstruction error is greater than a preset reconstruction error threshold, the sparsity is updated based on the optimal step factor, the to-be-used reconstruction data is determined based on the updated sparsity, and the to-be-used reconstruction data is determined as the target reconstruction data when the reconstruction error is less than the preset reconstruction error threshold.

[0052] It can be understood that if the reconstruction error is greater than the preset reconstruction error threshold, the reconstruction accuracy does not meet the standard, the optimal step factor can be found through the particle swarm optimization algorithm, the sparsity is updated, and then the signal is reconstructed based on the new sparsity and the observation value, until the reconstruction error of the to-be-used reconstruction data obtained finally meets the preset reconstruction error threshold, if it still does not meet the preset reconstruction error threshold, the optimal step factor is continuously determined and the sparsity is updated until the reconstruction error is less than the preset reconstruction error threshold, and the iteration is stopped.

[0053] On the basis, the historical reconstruction data is obtained, and the support set of the historical reconstruction data is determined.

[0054] The parameters of the measurement matrix are updated based on the support set, and the observation value is obtained based on the updated measurement matrix.

[0055] Specifically, the voltage, current and other data collected by the medium and low voltage equipment have strong correlation, the reconstruction information at the last moment is taken as prior information, the reconstruction efficiency and accuracy can be improved. Therefore, the historical reconstruction data needs to be obtained, and the corresponding support set needs to be determined. Further, the SP reconstruction algorithm has the characteristics of less measurement dimension and high reconstruction accuracy, and the backtracking idea is introduced, that is, whether the observation signal y exists in the current estimated value is judged by using the support set obtained by the last reconstruction, if not, the unusable elements are removed and the elements are updated accordingly.

[0056] The technical scheme of the embodiment of the application obtains the observation value corresponding to the power grid equipment state data based on the measurement matrix, the measurement matrix is a recovery error optimization measurement matrix, the observation value is processed based on the subspace pursuit algorithm, target reconstruction data corresponding to the power grid equipment state data is obtained through reconstruction, the existing compression sensing measurement matrix is random in observation time, the problem that the data amount is too redundant and the hardware resources are wasted exists besides that the sparsified signal cannot be effectively collected, the device feature information is retained to ensure a low reconstruction error of the data, the data collection amount is reduced, the communication overhead is effectively reduced, the data processing pressure of the data center is reduced, and the state monitoring and fault diagnosis capability of the medium and low voltage power distribution Internet of Things multi-source heterogeneous equipment is improved.

[0057] Embodiment Two

[0058] Figure 2 A process diagram of compressed sensing signal observation based on an adaptive measurement matrix is provided for Embodiment Two of the present application. This embodiment is a preferred embodiment of the above-mentioned embodiment, and its specific implementation can be referred to the technical solution of this embodiment. Among them, the same or corresponding technical terms as the above-mentioned embodiments will not be described here. As shown in the figure, the method comprises: Figure 2

[0059] The compressed sensing data observation and reconstruction algorithm based on the adaptive sensing matrix and the improved SP reconstruction algorithm proposed in the embodiment of the present application first optimizes the measurement matrix for state observation of the state information of the medium and low voltage power distribution Internet of Things device through the recovery error, and then uses the improved SP algorithm to reconstruct the signal, which preserves the device feature information and ensures a low reconstruction error of the data. This method greatly reduces the amount of data collection, effectively reduces the communication overhead, reduces the data processing pressure of the data center, and improves the state monitoring and fault diagnosis capability of the medium and low voltage power distribution Internet of Things multi-source heterogeneous device.

[0060] With the continuous advancement of the strong smart grid strategy of State Grid Corporation, it is required to monitor and collect information such as voltage, current, active / reactive power, real-time power consumption of users, switching quantity, and temperature in the station of devices such as RTU, DTU, FTU, TTU, and user collection terminals in medium and low voltage distribution network, in order to improve the level of power quality management and promote demand side response. The traditional compressed sensing collection strategy uses a pre-defined random matrix to collect high-frequency state information such as voltage and current, which has poor observation adaptability and consumes a large amount of hardware resources. At the same time, the OMP (Orthogonal Matching Pursuit) and SP (Subspace Pursuit) algorithms are used to directly reconstruct the collected data without considering the correlation between the data, resulting in low reconstruction accuracy and efficiency, which is difficult to meet the real-time monitoring demand of massive data in medium and low voltage distribution network.

[0061] The compressed sensing data collection method based on the adaptive measurement matrix and the improved SP reconstruction algorithm proposed in this embodiment reconstructs the collected voltage and current data in real time, obtains the unbiased estimated MSE as the objective function, and adjusts the sampling matrix according to the observation feedback result of the previous moment as the feedback result. At the same time, an improved SP algorithm is proposed, which uses the prior information estimated previously to more accurately reconstruct the sampling data of the next moment. The specific process is as follows:

[0062] 1) Adaptive measurement matrix generation algorithm objective function

[0063] ​The objective function of the present method is the mean square error (MSE) of the minimized recovery result of the voltage, current and power distribution system state data, i.e.

[0064]

[0065] wherein is the recovery result However, since the true situation of the medium and low voltage distribution network data is actually unknown during transmission, formula (1) is not applicable in practice. Accordingly, the present application draws on the idea of cognitive tracking radar, taking the MSE under the Cramer-Rao bound (CRB) as the target function for adaptive updating and adjustment of the observation matrix. Using the CRB, the minimum possible MSE of all unbiased estimators of the deterministic parameters can be obtained, and when the bound is very tight, reducing the CRB indicates that the MSE can be correspondingly reduced, and by pushing the CRB downward, the reconstruction performance of the voltage and current is significantly improved.

[0066] According to the results of the CRB on the recovery error of a sparse signal, given the sparsity ||x||0=K of the sparse vector, the constrained CRB is derived as formula (2)

[0067]

[0068] wherein tr(.) represents the trace of a matrix, Λ * represents the true support set containing the index of the non-zero elements in x, and (.) Λ represents the column vector of the matrix a in Λ.

[0069] When the signal-to-noise ratio (SNR) is large, the constrained CRB is very tight, which means that the noise is relatively small (the size of all non-zero elements is much larger than the standard deviation of the noise). Related literature discusses algorithms close to the boundary, and the SP method can be regarded as a greedy alternative algorithm to these algorithms. When a large number of independent samples are obtained, the CRB can be calculated using the maximum likelihood (ML) method as shown in formula (3).

[0070]

[0071] When the number of samples is infinite, the joint typicality (JT) estimator satisfies formula (2), provided that the sensing matrix is a Gaussian matrix with elements following Ν(0, 1 / N). However, the JT estimator has a large amount of calculation due to the exhaustive search of the support set. The SP calls the greedy idea to identify the support set, and then uses the least squares method to estimate the size of the index item in the support set.

[0072] 2) Adaptive measurement matrix generation and observation

[0073] The application minimizes the CRB as the objective function to generate the measurement matrix adaptively. According to the common method, we assume that each voltage and current sample x is independent and has a uniform form, ||φ m || 2 = 1, m = 1, 2, …, M. The goal of the adaptation is to have a better recovery result under the limited sampling power. For different application scenarios, the constraint can be in other forms. The whole sensing matrix Φ is divided into two parts. The first M0 rows of the measurement matrix are generated offline, and the remaining rows are constructed sequentially according to the previous measurement results.

[0074] In the initial step is a Gaussian matrix, and the elements are independent of each other and are identically distributed with N(0, 1 / N) regardless of x. The reason for using a Gaussian matrix is that such a matrix has a constant RIP, and the stability of the SP algorithm is guaranteed. In particular, M0>2K needs to satisfy the uniqueness condition.

[0075] In the next step, each row of is optimized sequentially to minimize the CRB, Since the true support set Λ * is not available, the estimated value Λ is used. Assuming that the measurement value at time instant m∈[M0,M-1], m is y 1:m = Φ 1:m x+e 1:m can be obtained, and the original signal is estimated according to the SP algorithm. The estimated value is denoted as The support set is denoted as Λ (m) . The list contained in Λ (m) is denoted as and The remaining part is denoted as and where (.) c denotes the complement. Therefore, after column rearrangement, the sampling matrix can be simplified to formula (4):

[0076]

[0077] Then φ m+1 is constructed by minimizing the CRB (4)

[0078]

[0079] where the variance σ 2 of the noise is ignored because it is independent of φ. Note that b T b≥0 and A is known, and using the matrix inverse formula, formula (4) can be further simplified to

[0080]

[0081] Let a T a = γ and

[0082]

[0083] Assume A T A is positive definite, and the objective function in (7) is monotonically increasing with respect to γ ∈ [0, 1]. The best γ = 1 and c = a, and (7) simplifies to (8).

[0084]

[0085] The solution to (8) is a opt = u min , and u min is the eigenvector of A T corresponding to the smallest eigenvalue. Because b opt = 0. Thus, the sampled vector φ m+1 is obtained, and the (m + 1)th observation can be made accordingly. The adaptively constructed sensing matrix helps improve the performance of the SP algorithm. The benefits come from two aspects: enhanced signal-to-noise ratio and mitigated interference among the original signals x.

[0086] The solution set φ m+1 indicates that the sampling resources are allocated according to Λ, and these elements are likely to be non-zero, which increases the signal-to-noise ratio. When there is a ready support set, the best approach is to focus on the components corresponding to Λ * and ignore the rest. If there is no such support set, the estimated support set is used and provides useful information.

[0087] The inconsistency among the columns in Λ mitigates the interference among the elements in Λ and helps improve the estimation. The criterion can be used to select the columns of A T A + aa T approximates the identity matrix. When it is an identity matrix, [A T , a] T are orthogonal to each other, avoiding the interference among the elements in x Λ . Let λ1, λ2, …, λ |Λ| denote the eigenvalues of the Gram matrix A T A + aa T , where ||·|| represents the cardinality of a set. Because A is known, and a T a = 1, the objective function is:

[0088]

[0089] Thus, the completed adaptive measurement matrix-based compressed sensing observation is completed, and the compressed observation of the state information of the multi-element power utilization equipment of the medium and low voltage distribution network is completed, and the specific process is as shown in Figure 2 Figure 2 The figure is a process diagram of the compressed sensing signal observation based on the adaptive measurement matrix.

[0090] 3) Improved SP reconstruction algorithm based on adaptive measurement matrix compressed sensing

[0091] The voltage, current and other data collected by the medium and low voltage equipment have strong correlation. The reconstruction information of the previous moment is used as prior information, which can improve the reconstruction efficiency and improve the reconstruction accuracy. The SP reconstruction algorithm is used to measure the dimension, which has the characteristics of less measurement and high reconstruction accuracy. The backtracking idea is introduced, that is, the support set obtained by the last reconstruction is used to judge whether the observation signal y exists in the current estimate value. If it does not exist, the unusable elements are removed and the elements are updated accordingly. The main process of the algorithm is 1) first, the sparsity of the signal is adaptively estimated according to the prior knowledge, 2) the columns of the observation matrix in the compressed sensing reconstruction are selected, 3) the reliability of the current subspace is calculated, the last column is removed, and a new selected column is added, and the sparsity is continued to be calculated. Until the iteration stopping condition is met, and the number of selected columns reaches the sparsity, it will not be expanded.

[0092] The flowchart of the improved SP algorithm is as shown in Figure 3 Figure 3 The figure is a process diagram of the improved SP reconstruction algorithm.

[0093] Meanwhile, the embodiment further proposes an improved SP reconstruction algorithm based on adaptive compressed sensing. The algorithm first adaptively estimates the sparsity K of the voltage and current data of the medium and low voltage distribution network equipment, denoted as k0, and then reconstructs the original signal in an iterative form, evaluates the reconstruction effect of the signal, and if the iteration termination condition is not met, the optimal step factor is found through the particle swarm optimization algorithm, the signal sparsity K is updated, and the signal reconstruction is continued. When the reconstruction error is less than the set threshold, it indicates that the reconstruction accuracy of the signal meets the requirements, and the iteration can be stopped. The reconstruction process of the method is as shown in Figure 4 Figure 4 The figure is a process diagram of the reconstruction of the embodiment.

[0094] The specific process of the reconstruction algorithm proposed in the embodiment is as follows:

[0095] (1) Estimate the sparsity of the original signal through the sparsity adaptive algorithm.

[0096] (2) Initialize the residual y r_n =y

[0097] (3) Obtain the support set of the signal​​​ K largest elements in the absolute value and iterate

[0098] (4) the indices K largest elements in the absolute value

[0099] (5) let where

[0100] (6) the set of indices of the K largest values

[0101] (7)

[0102] (8) if then exit the iteration, otherwise use the particle swarm optimization to obtain the optimal sparsity y r_n = y continue the iteration.

[0103] In this part, the superiority of the algorithm in this paper is verified by simulation experiments on matlab, and the method (adaptive SP, denoted as ASP), standard SP (SSP), Bayesian compressed sensing (BCS) and Finding Multiple Needles (FMN) algorithm are compared. Select the middle and low voltage distribution network current state signal as an example, select signals of different sizes respectively under the condition of compression ratio of 0.5, the MSE results of different compressed sensing measurement matrix construction methods are shown in Figure 5 , and the measurement error diagram of different measurement matrix construction algorithms is shown in Figure 5 .

[0104] As shown in Figure 5 , the FMN algorithm and the SSP algorithm have the same MSE when the sample number M = 64, and when the sample number increases, the performance of the ASP algorithm in this paper is better than that of the SSP and other discussed algorithms. Moreover, with the increase of the length of the current state signal, the reconstruction error becomes smaller and smaller, and the reconstruction effect is also better and better compared with other algorithms. At the same time, it is observed that the SP method has better performance in indicating the support set. This is because the adaptive scheme in ASP is more effective than BCS. Therefore, the algorithm in this paper is more suitable for the perception and monitoring of multi-element device state information in the massive data scenario of middle and low voltage distribution network.

[0105] The algorithm considers the practical performance while effectively reducing the data volume of the medium and low voltage power distribution Internet of Things state information transmission, ensures that the algorithm has high feasibility while effectively reducing communication pressure and cost and reducing the signal processing burden of the power system control center. Noise is a factor that greatly affects signal processing. Because the simulation platform is built on MATLAB in the present application, the state data compression transmission of the multi-source heterogeneous equipment of the medium and low voltage power distribution network is simulated, different degrees of white noise are set, and the noise resistance of the algorithm is tested. Finally, the algorithm is compared with other algorithms to compare the differences in noise resistance, and it is proved that the algorithm has better noise resistance under the same conditions. The hardware system for this test is Window 10 Professional, and the processor is Core i510300T. The results are shown in Figure 5 , Figure 6 The anti-noise performance diagram for actual deployment.

[0106] It can be seen from the comparison that the algorithm has excellent reconstruction accuracy under different degrees of noise interference, can effectively resist the influence of environmental noise on data acquisition, and has significantly better noise resistance than other algorithms, which can meet the real-time interaction demand of massive data of the medium and low voltage power distribution Internet of Things while ensuring the integrity of the data.

[0107] The technical scheme of the embodiment of the application obtains observation values corresponding to the state data of the power grid equipment based on a measurement matrix, the measurement matrix is a recovery error optimization measurement matrix, processes the observation values based on a subspace tracking algorithm, and reconstructs target reconstruction data corresponding to the state data of the power grid equipment, solves the problem that the existing compressed sensing measurement matrix is random in observation time, cannot effectively collect the sparse signal, and has excessive redundant data, wasting hardware resources, realizes low reconstruction error of the device feature information, reduces the data acquisition amount, effectively reduces the communication cost, reduces the data processing pressure of the data center, and improves the state monitoring and fault diagnosis capability of the multi-source heterogeneous equipment of the medium and low voltage power distribution Internet of Things.

[0108] Embodiment three

[0109] Figure 7 A structure diagram of a power grid equipment data processing device provided by the third embodiment of the application is shown in Figure 7 The device comprises:

[0110] An observation value determination module 310 is configured to obtain observation values corresponding to the state data of the power grid equipment based on a measurement matrix, and the measurement matrix is a recovery error optimization measurement matrix.

[0111] A data reconstruction module 320 is configured to process the observation values based on a subspace tracking algorithm and reconstruct target reconstruction data corresponding to the state data of the power grid equipment.

[0112] The technical scheme of the embodiment of the present application obtains observation values corresponding to power grid equipment state data based on a measurement matrix, the measurement matrix being a recovery error optimization measurement matrix; and the observation values are processed based on a subspace pursuit algorithm to obtain target reconstruction data corresponding to the power grid equipment state data, thereby solving the problem that the existing compressed sensing measurement matrix is random in observation time, cannot effectively collect signals after sparsification, and has excessive redundant data, wasting hardware resources, and achieving low reconstruction error of data by retaining equipment feature information, reducing data collection, effectively reducing communication overhead, reducing data processing pressure of a data center, and improving the state monitoring and fault diagnosis capability of a low-voltage power distribution Internet of Things multi-source heterogeneous device.

[0113] Optionally, the apparatus further comprises a fault judgment module,

[0114] determining operation data of the power grid equipment based on the target reconstruction data after the target reconstruction data corresponding to the power grid equipment state data is obtained;

[0115] If the operation data is consistent with preset fault data, it is determined that the power grid equipment is in a fault state.

[0116] Optionally, the observation value determination module 310 comprises:

[0117] a mean square error determination module configured to determine a mean square error function based on voltage data and current data of the power grid equipment and a Cramer-Rao bound criterion;

[0118] a measurement matrix generation module configured to determine the mean square error function as a target function, and generate the measurement matrix based on the target function and a Gaussian matrix.

[0119] Optionally, the data reconstruction module 320 comprises:

[0120] a sparsity determination module configured to predict sparsity of the power grid equipment state data based on the subspace pursuit algorithm;

[0121] a reconstruction error determination module configured to process the observation values based on the sparsity to obtain to-be-used reconstruction data, and determine reconstruction error of the to-be-used reconstruction data and the power grid equipment state data.

[0122] Optionally, the data reconstruction module 320 comprises:

[0123] a to-be-used reconstruction data determination module configured to determine the to-be-used reconstruction data as target reconstruction data when the reconstruction error of the to-be-used reconstruction data is less than a preset reconstruction error threshold.

[0124] Optionally, the data reconstruction module 320 comprises:

[0125] a sparsity updating module, configured to determine an optimal step factor corresponding to the to-be-used reconstruction data by using a particle swarm optimization algorithm if the reconstruction error is greater than a preset reconstruction error threshold, and update the sparsity based on the optimal step factor;

[0126] a target reconstruction data determining module, configured to determine the to-be-used reconstruction data based on the updated sparsity, and determine the to-be-used reconstruction data as target reconstruction data when the reconstruction error is less than the preset reconstruction error threshold.

[0127] Optionally, the apparatus further comprises:

[0128] a support set determining module, configured to acquire historical reconstruction data, and determine a support set of the historical reconstruction data;

[0129] an updating module, configured to update parameters of the measurement matrix based on the support set, and acquire the observation value based on the updated measurement matrix.

[0130] The processing apparatus for power grid equipment data provided in the embodiments of the present application can execute the processing method for power grid equipment data provided in any of the embodiments of the present application, and has the function modules and beneficial effects corresponding to the execution method.

[0131] Embodiment four

[0132] Figure 8 A structural schematic diagram of an electronic device is provided for embodiment four of the present application. The electronic device is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular telephones, smart phones, wearable devices (such as headsets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not intended to limit the implementations of the present application described and / or claimed in this document.

[0133] As Figure 8As shown, the electronic device 10 includes at least one processor 11, and a memory, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., connected to the at least one processor 11 in communication. The memory stores a computer program executable by the at least one processor 11, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or loaded into the random access memory (RAM) 13 from the storage unit 18. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0134] A plurality of components in the electronic device 10 are connected to the I / O interface 15, including an input unit 16, such as a keyboard, a mouse, etc., an output unit 17, such as various types of displays, a speaker, etc., a storage unit 18, such as a magnetic disk, an optical disk, etc., and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0135] The processor 11 can be various general-purpose and / or special-purpose processing components having processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 performs various methods and processes described above, such as the processing method of power grid device data.

[0136] In some embodiments, the processing method of power grid device data can be implemented as a computer program tangibly embodied in a computer readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the processing method of power grid device data described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to perform the processing method of power grid device data by any other appropriate means, such as by means of firmware.

[0137] The various embodiments of the systems and techniques described above can be implemented in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a load programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0138] Computer programs used to implement the processes of the application can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer program, when executed, can cause instructions defined in the flow charts and / or block diagrams to be implemented. The computer program can be executed entirely on a machine, partially on a machine, partially on a machine as a standalone software package and partially on a remote machine or entirely on a remote machine or server.

[0139] In the context of the present application, a computer-readable storage medium can be a tangible medium that can contain or store computer programs for use by or in connection with an instruction execution system, apparatus, or device. Computer-readable storage media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium will include one or more lines of electrical connections, portable computer disks, hard disk drives, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fibers, portable compact disc read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0140] To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0141] The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0142] The computing system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. A server can be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system, to solve the defects of large management difficulty and weak business scalability in traditional physical host and VPS service.

[0143] It should be understood that the various forms of flow shown above can be re-ordered, added to, or deleted from without departing from the scope of the present disclosure. For example, the steps recited in the present disclosure can be executed in parallel, executed in sequence, or executed in a different order, as long as the desired results of the present disclosure are achieved, and the present disclosure is not limited herein.

[0144] The specific embodiments described above are not intended to be limiting, and persons skilled in the art will appreciate that various modifications, combinations, sub-combinations and alternatives can be made to the specific embodiments without departing from the spirit and principles of the disclosure. Accordingly, the disclosure is not limited to the specific embodiments described above, but only by the scope of the appended claims.

Claims

1. A method for processing data from power grid equipment, characterized in that, include: The observations corresponding to the power grid equipment status data are obtained based on the measurement matrix, which is a recovery error optimized measurement matrix; The observations are processed using a subspace tracking algorithm to reconstruct target reconstruction data corresponding to the power grid equipment status data. The step of acquiring observations corresponding to the state data of power grid equipment based on a measurement matrix, wherein the measurement matrix is ​​a recovery error optimized measurement matrix, includes: determining a mean square error function based on the voltage data, current data and Cramer-Rhodes bound criterion of the power grid equipment; determining the mean square error function as an objective function; and generating the measurement matrix based on the objective function and a Gaussian matrix. The process of processing the observations based on the subspace tracking algorithm to reconstruct target reconstructed data corresponding to the power grid equipment status data includes: predicting the sparsity of the power grid equipment status data based on a sparsity adaptive algorithm; processing the observations based on the sparsity to obtain reconstructed data to be used, and determining the reconstruction error between the reconstructed data to be used and the power grid equipment status data; When the reconstruction error of the reconstruction data to be used is less than a preset reconstruction error threshold, the reconstruction data to be used is determined as the target reconstruction data; If the reconstruction error is greater than a preset reconstruction error threshold, the optimal step size factor corresponding to the reconstruction data to be used is determined by the particle swarm optimization algorithm, and the sparsity is updated based on the optimal step size factor; the reconstruction data to be used is determined based on the updated sparsity, until the reconstruction error is less than the preset reconstruction error threshold, and then the reconstruction data to be used is determined as the target reconstruction data. The specific reconstructed target data corresponding to the power grid equipment status data obtained from the reconstruction includes: (1) Estimate the sparsity of the original signal using a sparsity adaptive algorithm; (2) Initialize the residual y r_n =y (3) Obtain the support set of the signal Find the K largest elements in the set and calculate the residuals. and iterate (4) Subscript K indices of the element with the largest absolute value (5) Order in (6) The set of the following tables containing the K maximum values. (7) (8) If but Exit the iteration; otherwise, use particle swarm optimization to obtain the optimal sparsity y. r_n =y continues to iterate.

2. The method according to claim 1, characterized in that, After the target reconstructed data corresponding to the power grid equipment status data is obtained through reconstruction, the method further includes: The operating data of the power grid equipment are determined based on the target reconstruction data; If the operating data is consistent with the preset fault data, then the power grid equipment is determined to be in a fault state.

3. The method according to claim 1, characterized in that, Also includes: Acquire historical reconstruction data and determine the support set of the historical reconstruction data; The parameters of the measurement matrix are updated based on the support set, and the observation values ​​are obtained based on the updated measurement matrix.

4. A data processing device for power grid equipment, characterized in that, include: The observation determination module is used to obtain the observations corresponding to the power grid equipment status data based on the measurement matrix, wherein the measurement matrix is ​​a recovery error optimized measurement matrix; The data reconstruction module is used to process the observations based on the subspace tracking algorithm to reconstruct the target reconstruction data corresponding to the power grid equipment status data; The observation determination module includes: The mean square error determination module is used to determine the mean square error function based on voltage and current data of power grid equipment and the Cramer-Rhodes boundary criterion. The measurement matrix generation module is used to determine the mean square error function as the objective function, and generate the measurement matrix based on the objective function and the Gaussian matrix. The data reconstruction module includes: A sparsity determination module is used to predict the sparsity of the power grid equipment status data based on a sparsity adaptive algorithm. The reconstruction error determination module is used to process the observations based on the sparsity to obtain the reconstructed data to be used, and to determine the reconstruction error between the reconstructed data to be used and the power grid equipment status data. The module for determining the reconstructed data to be used is used to determine the reconstructed data to be used as the target reconstructed data when the reconstruction error of the reconstructed data to be used is less than a preset reconstruction error threshold. The sparsity update module is used to determine the optimal step size factor corresponding to the reconstruction data to be used through particle swarm optimization algorithm if the reconstruction error is greater than a preset reconstruction error threshold, and update the sparsity based on the optimal step size factor. The target reconstruction data determination module is used to determine the reconstruction data to be used based on the updated sparsity, until the reconstruction error is less than the preset reconstruction error threshold, and then determine the reconstruction data to be used as the target reconstruction data. The specific reconstructed target data corresponding to the power grid equipment status data obtained from the reconstruction includes: (1) Estimate the sparsity of the original signal using a sparsity adaptive algorithm; (2) Initialize the residual y r_n =y (3) Obtain the support set of the signal Find the K largest elements in the set and calculate the residuals. and iterate (4) Subscript K indices of the element with the largest absolute value (5) Order in (6) The set of the following tables containing the K maximum values. (7) (8) If but Exit the iteration; otherwise, use particle swarm optimization to obtain the optimal sparsity y. r_n =y continues to iterate.

5. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the method for processing power grid equipment data according to any one of claims 1-3.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the method for processing power grid equipment data as described in any one of claims 1-3.

Images