A method and system for online monitoring of power grid switch operation status

Abstract

The present invention relates to the technical field of power grid switch monitoring, and discloses a method and system for online monitoring of the operating status of a power grid switch, the method comprising: collecting and preprocessing the operating parameter data of a power grid switch cabinet; performing dimensionality reduction processing on the preprocessed operating parameter data of the power grid switch cabinet; inputting the low-dimensional power grid switch cabinet operating parameter data into a power grid switch cabinet operating status recognition model to obtain the operating status of the power grid switch cabinet and perform abnormal monitoring. The present invention achieves data dimensionality reduction for probability distribution similarity while ensuring that the distribution structure of the power grid switch cabinet operating parameter data remains unchanged, and uses the singular value sequence of the low-dimensional mapping matrix as a hidden variable to characterize the scale, energy and principal component of the low-dimensional power grid switch cabinet operating parameter data, and performs multi-step convolution processing on the hidden variable to obtain the characteristic information of the hidden variable at the dynamic step scale, which is used as a dynamic hidden variable to characterize the short-time and multi-scale transformation information of the power grid switch cabinet for abnormal monitoring.

Description

Technical Field

[0001] The present invention relates to the field of power grid switch monitoring, and in particular to an online monitoring method and system for the operation status of a power grid switch. Background Art

[0002] With the development of modern society, the power system, as the lifeline of economic and social development, undertakes the important task of providing stable and reliable electric energy for all walks of life. In the power system, the grid switch is one of the core components of the operation and control of power equipment, and plays a key role in protecting and controlling the power system. Once these switchgear fail, it will not only affect the normal transmission of electricity, but also may cause large-scale power outages, and even cause equipment damage and casualties. Therefore, it is of great significance to monitor the status of the grid switch in real time, detect potential faults early, and take maintenance measures in time to ensure the safe and stable operation of the power system. Traditional grid switch status monitoring mainly relies on regular manual inspection and maintenance. However, with the expansion of the scale and complexity of the power system, this method has been difficult to meet the requirements of modern power systems for real-time and accuracy. First, manual inspection has obvious lag, which may cause the switchgear to fail to be discovered in time between two inspections. Secondly, the cost of manual maintenance is high, especially in large-scale power systems, and regular maintenance of each switch requires a lot of manpower and material resources. In addition, the traditional regular maintenance mode cannot achieve dynamic adjustment of the switch operation status, which easily leads to waste or insufficient maintenance resources. Summary of the invention

[0003] In view of this, the present invention provides an online monitoring method for the operating status of a power grid switch, which realizes in-depth prediction of the switch operating status by real-time collection of power grid switch operating data for big data analysis, thereby achieving the purpose of precise maintenance and optimal allocation of resources.

[0004] To achieve the above object, the present invention provides a method for online monitoring of the operating status of a power grid switch, comprising the following steps:

[0005] S1: collecting and preprocessing the operating parameter data of the power grid switch cabinet to obtain the preprocessed operating parameter data of the power grid switch cabinet, wherein the multi-stage noise reduction is an implementation method of the preprocessing, wherein the operating parameters of the power grid switch cabinet include current, voltage, power, insulation resistance, power grid frequency, harmonic content, temperature and humidity;

[0006] S2: Performing dimensionality reduction processing on the pre-processed power grid switch cabinet operating parameter data to obtain low-dimensional power grid switch cabinet operating parameter data;

[0007] S3: constructing a power grid switch cabinet operation status identification model, wherein the power grid switch cabinet operation status identification model takes low-dimensional power grid switch cabinet operation parameter data as input and takes power grid switch cabinet operation status as output, wherein dynamic latent variable analysis is an implementation method of the power grid switch cabinet operation status identification model;

[0008] S4: Input the low-dimensional grid switch cabinet operating parameter data into the grid switch cabinet operating state recognition model to obtain the grid switch cabinet operating state and perform abnormal monitoring. If an abnormality is detected, an alarm is processed.

[0009] As a further improvement method of the present invention:

[0010] Optionally, collecting the operating parameter data of the power grid switch cabinet in step S1 includes:

[0011] Collect the operating parameter data of the power grid switch cabinet, where the operating parameters of the power grid switch cabinet include current, voltage, power, insulation resistance, power grid frequency, harmonic content, temperature and humidity. The collected operating parameter data are:

[0012]

[0013] in:

[0014] x represents the operating parameter data;

[0015] x n The data sequence representing the nth operating parameter index, N represents the total number of operating parameter indexes, wherein the first to the Nth operating parameter indexes are current, voltage, power, insulation resistance, grid frequency, harmonic content, temperature and humidity;

[0016] Represents the data sequence x n The data values ​​at M sampling moments, Represents the data sequence x n The data value at the mth sampling moment in , m∈[1,M], M represents the number of discrete data values ​​collected in real time during the real-time operation parameter data collection process;

[0017] The operating parameter data x is preprocessed, and the multi-stage noise reduction is an implementation method of the preprocessing to obtain the preprocessed power grid switch cabinet operating parameter data X.

[0018] Optionally, the preprocessing of the operating parameter data x includes:

[0019] The operating parameter data x is preprocessed, wherein the preprocessing process is:

[0020] S11: Convert any data sequence in the operating parameter data x into a matrix form, where the data sequence xn The corresponding matrix is:

[0021]

[0022] in:

[0023] E n Represents the data sequence x n The corresponding matrix,

[0024] S12: Perform singular value decomposition on the transformed matrix to obtain the left and right singular matrices and the singular value diagonal matrix of the matrix, where the matrix E n The left singular matrix of The right singular matrix is The singular value diagonal matrix is ​​∑ n :

[0025]

[0026] in:

[0027] Represents the matrix E n a singular values ​​of , diag(·) indicates diagonal matrixing;

[0028] S13: Perform spectral representation on any data sequence in the operating parameter data x to obtain the spectral representation result of the data sequence at a number of transformation points, where Data sequence x n The spectrum representation result at the i-th transformation point is:

[0029]

[0030] in:

[0031] Represents the data sequence x n The spectrum representation result at the i-th transformation point;

[0032] exp(·) represents an exponential function with a natural constant as the base;

[0033] j represents the imaginary unit;

[0034] S14: Perform difference change on the spectrum representation result to obtain a difference sequence of the data sequence at a transformation points, where the data sequence x n The corresponding difference sequence is:

[0035]

[0036] in:

[0037] Pn Represents the data sequence x n The corresponding difference sequence is The spectrum shows the result and The energy difference between

[0038] S15: Construct the period threshold of the data sequence, where the data sequence x n The corresponding cycle threshold is:

[0039]

[0040] in:

[0041] T n Represents the data sequence x n The corresponding cycle threshold;

[0042] σ n Represents the difference sequence P n The standard deviation of

[0043] μ n Represents the difference sequence P n The mean of

[0044] S16: The number of singular values ​​required to generate the data sequence, where the data sequence x n The number of singular values ​​to be retained is num n :

[0045]

[0046] in:

[0047] represents an intermediate variable;

[0048] S17: Filter the singular value diagonal matrix based on the number of singular values ​​and form a denoising matrix of the data sequence, where the data sequence x n The corresponding denoising matrix is:

[0049]

[0050] in:

[0051] T stands for transpose;

[0052] E′ n Represents the data sequence x n The corresponding denoising matrix, the data distribution form of the denoising matrix is ​​the same as the matrix E n Consistency;

[0053] ∑′ n Represents a diagonal matrix ∑ that retains singular values nMiddle front num n singular values ​​and set the remaining singular values ​​to 0;

[0054] S18: According to the inverse operation of converting the data sequence into a matrix, a data sequence of length M is extracted from the denoising matrix, and the extracted data sequence is normalized to obtain a normalized data sequence of N operating parameter indicators, wherein the normalized data sequence of the nth operating parameter indicator is:

[0055]

[0056] in:

[0057] X n Represents the normalized data sequence of the nth operating parameter indicator; Represents the normalized data sequence X n The data value at the mth sampling moment in ;

[0058] The pre-processed grid switch cabinet operating parameter data X = {X n |n∈[1,N]}.

[0059] Optionally, in step S2, the preprocessed power grid switch cabinet operating parameter data is subjected to dimensionality reduction processing, including:

[0060] The pre-processed power grid switch cabinet operating parameter data X is subjected to dimensionality reduction processing to obtain low-dimensional power grid switch cabinet operating parameter data, wherein the dimensionality reduction processing flow is as follows:

[0061] S21: Convert the pre-processed grid switch cabinet operating parameter data X into a sequence form:

[0062] X′=(X′(1),X′(2),...,X′(m),...,X′(M))

[0063]

[0064] in:

[0065] X′ represents the sequence representation of the preprocessed power grid switchgear operating parameter data X;

[0066] X′(m) represents the operating parameter data sequence at the mth sampling moment;

[0067] S22: Calculate the conditional probability that different operating parameter data sequences are similar, where the operating parameter data sequence X′(

[0068] m 1 ) and X′(m 2 )The similar conditional probability is:

[0069]

[0070] in:

[0071] p(X′(m 1 ); X′(m 2 )) represents the operating parameter data sequence X′(m 1 ) and X′(m 2 ) similar conditional probability, m 1 ,m 2 ∈

[0072] [1,M],m 1 ≠m 2 ;

[0073] dis(X′(m 1 ),X′(m 2 )) represents the operating parameter data sequence X′(m 1 ) and X′(m 2 ) between the two sides;

[0074] δ(m 1 ) represents the operating parameter data sequence X′(m 1 )’s standard deviation;

[0075] W n Indicates the indicator weight of the nth operating parameter indicator;

[0076] T n Indicates the data sequence x corresponding to the nth operating parameter indicator n The cycle threshold of

[0077] S23: Initialize and generate low-dimensional structure data L 0 =(L 0 (1),L 0 (2),...,L 0 (m),...,L 0 (M)), where L 0 (m) is the low-dimensional representation sequence of the operating parameter data sequence X′(m);

[0078] Set the low-dimensional structure data L 0 The current number of iterations is t, the maximum number of iterations is Max, and the initial value of t is 0;

[0079] S24: Calculate the conditional probability of similarity between different low-dimensional representation sequences in the low-dimensional structure data, and iterate the low-dimensional structure data in combination with the conditional probability of similarity between the operating parameter data sequences, and obtain the low-dimensional power grid switch cabinet operating parameter data L = (L(1), L(2), ..., L(m), ..., L(M)) corresponding to the preprocessed power grid switch cabinet operating parameter data X, where L(m) is the low-dimensional power grid switch cabinet operating parameter representation sequence corresponding to the operating parameter data sequence X′(m).

[0080] Optionally, in step S24, the conditional probabilities of similarity of different low-dimensional representation sequences in the low-dimensional structure data are calculated, and the low-dimensional structure data is iterated in combination with the conditional probabilities of similarity of the operating parameter data sequences:

[0081] S241: Calculate the low-dimensional structure data L obtained at the tth iteration t The conditional probability that any two sets of low-dimensional representation sequences are similar, where L t (m 1 ) and L t (m 2 )The similar conditional probability is:

[0082]

[0083] in:

[0084] q(L t (m 1 ),L t (m 2 )) indicates L t (m 1 ) and L t (m 2 ) similar conditional probabilities;

[0085] ||·|| represents the L1 norm;

[0086] S242: Constructing the objective function of low-dimensional structure data optimization:

[0087]

[0088] in:

[0089] F(·) represents the objective function, F(L t ) represents the low-dimensional structure data L t The corresponding objective function value;

[0090] p(X′(m 1 ),X′(m 2 )) represents X′(m 1 ),X′(m 2 ) are similar joint probabilities;

[0091] Low-dimensional structured data L t The smaller the corresponding objective function value is, the lower the dimensional structure data L is. t The more consistent the similarity distribution is with the similarity distribution of the grid switchgear operating parameter data;

[0092] S243: Taking minimizing the objective function as the optimization goal, the low-dimensional structure data L t Iteration is performed, where the iteration formula is:

[0093]

[0094] in:

[0095] η t represents the iteration weight coefficient;

[0096] represents the gradient of the objective function;

[0097] v t represents the iterative momentum, β represents the momentum weight coefficient; in the embodiment of the present invention, β is set to 0.9;

[0098] S244: Let t=t+1, and return to step S241 until the maximum number of iterations is reached, and the low-dimensional structure data obtained by the iteration at this time is used as the low-dimensional power grid switch cabinet operation parameter data L.

[0099] Optionally, constructing a power grid switch cabinet operating state identification model in step S3 includes:

[0100] Constructing a power grid switch cabinet operation state identification model, wherein dynamic latent variable analysis is an implementation method of the power grid switch cabinet operation state identification model, the power grid switch cabinet operation state identification model takes low-dimensional power grid switch cabinet operation parameter data as input and takes power grid switch cabinet operation state as output, wherein the power grid switch cabinet operation state identification model includes an input layer, a latent variable extraction layer, a dynamic representation layer and an identification layer;

[0101] The input layer is used to receive low-dimensional grid switchgear operating parameter data;

[0102] The latent variable extraction layer is used to extract latent variables from the low-dimensional grid switchgear operating parameter data;

[0103] The dynamic representation layer is used to dynamically represent the latent variables and obtain the dynamic latent variables of the low-dimensional power grid switchgear operating parameter data;

[0104] The recognition layer is used to identify the states of dynamic latent variables, where the states include sleep, normal operation, and abnormality.

[0105] Optionally, in step S4, the low-dimensional power grid switch cabinet operating parameter data is input into a power grid switch cabinet operating state recognition model to obtain the power grid switch cabinet operating state and perform abnormal monitoring, including:

[0106] The low-dimensional grid switchgear operating parameter data is input into the grid switchgear operating state recognition model to obtain the grid switchgear operating state, wherein the grid switchgear operating state recognition process is as follows:

[0107] S41: The input layer receives low-dimensional power grid switch cabinet operating parameter data L = (L(1), L(2), ..., L(m), ..., L(M));

[0108] S42: The latent variable extraction layer extracts the latent variable γ in the low-dimensional power grid switchgear operating parameter data L;

[0109] S43: The dynamic characterization layer dynamically characterizes the latent variable γ to obtain the dynamic latent variable Y of the low-dimensional power grid switchgear operating parameter data, wherein the dynamic characterization formula is:

[0110]

[0111] in:

[0112] Q h represents the convolution matrix of the hth scale, H represents the maximum scale, and * represents the convolution operator; the convolution steps of convolution matrices of different scales are different;

[0113] S44: The recognition layer performs state recognition on the dynamic latent variable Y, where the state recognition formula is:

[0114]

[0115] in:

[0116] State indicates the operating status of the power grid switch cabinet, where ―1, 0, and 1 correspond to abnormal, dormant, and normal operating states respectively;

[0117] Q θ Represents the mapping matrix corresponding to the state θ, and selects The maximum state is taken as the operating state of the power grid switch cabinet, and an alarm is issued if an abnormality is detected.

[0118] Optionally, in step S42, the latent variable extraction layer extracts latent variables in the low-dimensional power grid switch cabinet operating parameter data L, including:

[0119] S421: Generate low-dimensional parameter matrix S:

[0120] S=L T L

[0121] S422: Randomly generate mapping coefficient α 0 , and set the current iteration number of the mapping coefficient to d, the maximum iteration number to D, then the dth iteration result of the mapping coefficient is α d , the initial value of d is 0;

[0122] And calculate the weight coefficient w 0 :

[0123] w 0 =[w 0 (1),w 0 (2),...,w 0 (m),...,w 0 (M)]

[0124]

[0125] in:

[0126] w 0 (m) represents the weight of L(m), ||·|| 2 represents the L2 norm;

[0127] Then the dth iteration result of the weight coefficient is w d ;

[0128] S423: Generate mapping coefficient α d With the weight coefficient w d The corresponding intermediate parameter matrix:

[0129]

[0130] in:

[0131] S 1 (α d ) represents the mapping coefficient α d The corresponding intermediate parameter matrix, S 2 (w d ) represents the weight coefficient w d The corresponding intermediate parameter matrix;

[0132] An operator representing a tensor product computation;

[0133] I represents the identity matrix;

[0134] S424: Iterate the mapping coefficients and weight coefficients:

[0135]

[0136] The iterative error term of the mapping coefficient is calculated, where α d+1 The iterative error term is ρ(αd+1 ):

[0137]

[0138] S425: If the iterative error term is less than the preset threshold, α d+1 as the mapping coefficient, otherwise set d=d+1 and return to step S424;

[0139] S426: Based on the mapping coefficient α d+1 , calculate the low-dimensional mapping matrix S' corresponding to the low-dimensional parameter matrix S:

[0140]

[0141] S427: Perform singular value decomposition on the low-dimensional mapping matrix S′, and use the sequence of M singular values ​​obtained by the decomposition as the latent variable γ = (γ(1), γ(2), ..., γ(m), ..., γ(M)), where γ(m) represents the mth singular value obtained by the decomposition.

[0142] In order to solve the above problems, the present invention further provides an online monitoring system for the operation status of a power grid switch, characterized in that the system comprises:

[0143] The data acquisition module is used to collect the operating parameter data of the power grid switch cabinet and perform preprocessing to obtain the preprocessed operating parameter data of the power grid switch cabinet;

[0144] A dimension reduction processing module is used to perform dimension reduction processing on the pre-processed power grid switch cabinet operating parameter data to obtain low-dimensional power grid switch cabinet operating parameter data;

[0145] The state monitoring device is used to construct a power grid switch cabinet operation state recognition model, input low-dimensional power grid switch cabinet operation parameter data into the power grid switch cabinet operation state recognition model, obtain the power grid switch cabinet operation state and perform abnormal monitoring, and perform alarm processing if an abnormality is detected.

[0146] In order to solve the above problem, the present invention further provides an electronic device, the electronic device comprising:

[0147] A memory storing at least one instruction;

[0148] Communication interface, enabling electronic equipment to communicate; and

[0149] The processor executes the instructions stored in the memory to implement the above-mentioned method for online monitoring of the operating status of the power grid switch.

[0150] In order to solve the above problem, the present invention also provides a computer-readable storage medium, in which at least one instruction is stored. The at least one instruction is executed by a processor in an electronic device to implement the above-mentioned online monitoring method for the operating status of a power grid switch.

[0151] Compared with the prior art, the present invention proposes an online monitoring method for the operation status of a power grid switch, which has the following advantages:

[0152] First, this scheme collects the operating parameter data of the power grid switch cabinet and performs multi-level analysis and noise reduction on the collected data, including singular value decomposition, spectrum decomposition, etc. The decomposition results are filtered according to the coefficient of variation of the data sequence and the ratio of the spectrum energy difference to obtain the operating parameter data of the power grid switch cabinet with the power grid environment noise and system data fluctuations removed. The weighted joint probability of the similarity of the operating parameter data at different times and the probability of the low-dimensional data corresponding to the operating parameter data are similar. While ensuring that the distribution structure of the power grid switch cabinet operating parameter data remains unchanged, the momentum parameter is combined to perform rapid iterative processing of the low-dimensional data to obtain low-dimensional power grid switch cabinet operating parameter data with similar probability distribution, thereby reducing the computing resources required for subsequent operating state identification.

[0153] At the same time, this scheme combines the weight coefficients of parameter data at different times to update and iterate the mapping coefficients extracted from the latent variables, and uses the mapping coefficients to extract the low-dimensional mapping matrix from the low-dimensional power grid switchgear operating parameter data. The singular value sequence of the low-dimensional mapping matrix is ​​used as the latent variable to characterize the scale, energy and principal components of the low-dimensional power grid switchgear operating parameter data, and the latent variables are subjected to multi-step convolution processing to obtain the characteristic information of the latent variables at the dynamic step scale, which is used as the dynamic latent variable to characterize the short-time and multi-scale transformation information of the power grid switchgear, and the operating status of the power grid switchgear is monitored for abnormalities. BRIEF DESCRIPTION OF THE DRAWINGS

[0154] Figure 1 A schematic diagram of a flow chart of a method for online monitoring of the operating status of a power grid switch provided by an embodiment of the present invention;

[0155] Figure 2 A functional module diagram of a system for online monitoring of the operation status of a power grid switch provided by an embodiment of the present invention;

[0156] Figure 2 In: 100 online monitoring system for operation status of power grid switch, 101 data acquisition module, 102 dimension reduction processing module, 103 status monitoring device;

[0157] Figure 3 A schematic diagram of the structure of an electronic device for implementing a method for online monitoring of the operating status of a power grid switch provided by an embodiment of the present invention.

[0158] Figure 3 In: 1 electronic device, 10 processor, 11 memory, 12 program, 13 communication interface;

[0159] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0160] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0161] The embodiment of the present application provides an online monitoring method for the operating status of a power grid switch. The execution subject of the online monitoring method for the operating status of a power grid switch includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present application. In other words, the online monitoring method for the operating status of a power grid switch can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0162] Embodiment 1:

[0163] S1: Collecting and preprocessing the operating parameter data of the power grid switch cabinet to obtain the preprocessed operating parameter data of the power grid switch cabinet, wherein the operating parameters of the power grid switch cabinet include current, voltage, power, insulation resistance, power grid frequency, harmonic content, temperature and humidity.

[0164] The operation parameter data of the power grid switch cabinet is collected in step S1, including:

[0165] Collect the operating parameter data of the power grid switch cabinet, where the operating parameters of the power grid switch cabinet include current, voltage, power, insulation resistance, power grid frequency, harmonic content, temperature and humidity. The collected operating parameter data are:

[0166]

[0167] in:

[0168] x represents the operating parameter data;

[0169] x n The data sequence representing the nth operating parameter index, N represents the total number of operating parameter indexes, wherein the first to the Nth operating parameter indexes are current, voltage, power, insulation resistance, grid frequency, harmonic content, temperature and humidity;

[0170] Represents the data sequence x n The data values ​​at M sampling moments, Represents the data sequence x n The data value at the mth sampling moment in , m∈[1,M], M represents the number of discrete data values ​​collected in real time during the real-time operation parameter data collection process;

[0171] The operating parameter data x is preprocessed to obtain preprocessed power grid switch cabinet operating parameter data X.

[0172] The preprocessing of the operating parameter data x includes:

[0173] The operating parameter data x is preprocessed, wherein the preprocessing process is:

[0174] S11: Convert any data sequence in the operating parameter data x into a matrix form, where the data sequence x n The corresponding matrix is:

[0175]

[0176] in:

[0177] E n Represents the data sequence x n The corresponding matrix,

[0178] S12: Perform singular value decomposition on the transformed matrix to obtain the left and right singular matrices and the singular value diagonal matrix of the matrix, where the matrix E n The left singular matrix of The right singular matrix is The singular value diagonal matrix is ​​∑ n :

[0179]

[0180] in:

[0181] Represents the matrix E n a singular values ​​of , diag(·) indicates diagonal matrixing;

[0182] S13: Perform spectral representation on any data sequence in the operating parameter data x to obtain the spectral representation result of the data sequence at a number of transformation points, where Data sequence x n The spectrum representation result at the i-th transformation point is:

[0183]

[0184] in:

[0185] Represents the data sequence xn The spectrum representation result at the i-th transformation point;

[0186] exp(·) represents an exponential function with a natural constant as the base;

[0187] j represents the imaginary unit;

[0188] S14: Perform difference change on the spectrum representation result to obtain a difference sequence of the data sequence at a transformation points, where the data sequence x n The corresponding difference sequence is:

[0189]

[0190] in:

[0191] P n Represents the data sequence x n The corresponding difference sequence is The spectrum shows the result and The energy difference between

[0192] S15: Construct the period threshold of the data sequence, where the data sequence x n The corresponding cycle threshold is:

[0193]

[0194] in:

[0195] T n Represents the data sequence x n The corresponding cycle threshold;

[0196] σ n Represents the difference sequence P n The standard deviation of

[0197] μ n Represents the difference sequence P n The mean of

[0198] S16: The number of singular values ​​required to generate the data sequence, where the data sequence x n The number of singular values ​​to be retained is num n :

[0199]

[0200] in:

[0201] represents an intermediate variable;

[0202] S17: Filter the singular value diagonal matrix based on the number of singular values ​​and form a denoising matrix of the data sequence, where the data sequence x n The corresponding denoising matrix is:

[0203]

[0204] in:

[0205] T stands for transpose;

[0206] E′ n Represents the data sequence x n The corresponding denoising matrix, the data distribution form of the denoising matrix is ​​the same as the matrix E n Consistency;

[0207] ∑′ n Represents a diagonal matrix Σ that retains singular values n Middle front num n singular values ​​and set the remaining singular values ​​to 0;

[0208] S18: According to the inverse operation of converting the data sequence into a matrix, a data sequence of length M is extracted from the denoising matrix, and the extracted data sequence is normalized to obtain a normalized data sequence of N operating parameter indicators, wherein the normalized data sequence of the nth operating parameter indicator is:

[0209]

[0210] in:

[0211] X n Represents the normalized data sequence of the nth operating parameter indicator; Represents the normalized data sequence X n The data value at the mth sampling moment in ;

[0212] The pre-processed grid switch cabinet operating parameter data X = {X n |n∈[1,N]}.

[0213] S2: Perform dimensionality reduction processing on the pre-processed power grid switch cabinet operating parameter data to obtain low-dimensional power grid switch cabinet operating parameter data.

[0214] In the step S2, the pre-processed grid switch cabinet operating parameter data is subjected to dimensionality reduction processing, including:

[0215] The pre-processed power grid switch cabinet operating parameter data X is subjected to dimensionality reduction processing to obtain low-dimensional power grid switch cabinet operating parameter data, wherein the dimensionality reduction processing flow is as follows:

[0216] S21: Convert the pre-processed grid switch cabinet operating parameter data X into a sequence form:

[0217] X′=(X′(1),X′(2),...,X′(m),...,X′(<))

[0218]

[0219] in:

[0220] X′ represents the sequence representation of the preprocessed power grid switchgear operating parameter data X;

[0221] X′(m) represents the operating parameter data sequence at the mth sampling moment;

[0222] S22: Calculate the conditional probability that different operating parameter data sequences are similar, where the operating parameter data sequence X′(

[0223] m 1 ) and X′(m 2 )The similar conditional probability is:

[0224]

[0225] in:

[0226] p(X′(m 1 ); X′(m 2 )) represents the operating parameter data sequence X′(m 1 ) and X′(m 2 ) similar conditional probability, m 1 ,m 2 ∈

[0227] [1,M],m 1 ≠m 2 ;

[0228] dis(X′(m 1 ),X′(m 2 )) represents the operating parameter data sequence X′(m 1 ) and X′(m 2 ) between the two sides;

[0229] δ(m 1 ) represents the operating parameter data sequence X′(m 1 )’s standard deviation;

[0230] W n Indicates the indicator weight of the nth operating parameter indicator;

[0231] T n Indicates the data sequence x corresponding to the nth operating parameter indicatorn The cycle threshold of

[0232] S23: Initialize and generate low-dimensional structure data L 0 =(L 0 (1),L 0 (2),...,L 0 (m),...,L 0 (M)), where L 0 (m) is the low-dimensional representation sequence of the operating parameter data sequence X′(m);

[0233] Set the low-dimensional structure data L 0 The current number of iterations is t, the maximum number of iterations is Max, and the initial value of t is 0;

[0234] S24: Calculate the conditional probability of similarity between different low-dimensional representation sequences in the low-dimensional structure data, and iterate the low-dimensional structure data in combination with the conditional probability of similarity between the operating parameter data sequences, and obtain the low-dimensional power grid switch cabinet operating parameter data L = (L(1), L(2), ..., L(m), ..., L(M)) corresponding to the preprocessed power grid switch cabinet operating parameter data X, where L(m) is the low-dimensional power grid switch cabinet operating parameter representation sequence corresponding to the operating parameter data sequence X′(m).

[0235] In the step S24, the conditional probabilities of similarity of different low-dimensional representation sequences in the low-dimensional structure data are calculated, and the conditional probabilities of similarity of the operating parameter data sequences are combined to iterate the low-dimensional structure data:

[0236] S241: Calculate the low-dimensional structure data L obtained at the tth iteration t The conditional probability that any two sets of low-dimensional representation sequences are similar, where L t (m 1 ) and L t (m 2 )The similar conditional probability is:

[0237]

[0238] in:

[0239] q(L t (m 1 ),L t (m 2 )) indicates L t (m 1 ) and L t (m 2 ) similar conditional probabilities;

[0240] ||·|| represents the L1 norm;

[0241] S242: Constructing the objective function of low-dimensional structure data optimization:

[0242]

[0243] in:

[0244] F(·) represents the objective function, F(L t ) represents the low-dimensional structure data L t The corresponding objective function value;

[0245] p(X′(m 1 ),X′(m 2 )) represents X′(m 1 ),X′(m 2 ) are similar joint probabilities;

[0246] Low-dimensional structured data L t The smaller the corresponding objective function value is, the lower the dimensional structure data L is. t The more consistent the similarity distribution is with the similarity distribution of the grid switchgear operating parameter data;

[0247] S243: Taking minimizing the objective function as the optimization goal, the low-dimensional structure data L t Iteration is performed, where the iteration formula is:

[0248]

[0249] in:

[0250] η t represents the iteration weight coefficient;

[0251] represents the gradient of the objective function;

[0252] v t represents the iterative momentum, β represents the momentum weight coefficient; in the embodiment of the present invention, β is set to 0.9;

[0253] S244: Let t=t+1, and return to step S241 until the maximum number of iterations is reached, and the low-dimensional structure data obtained by the iteration at this time is used as the low-dimensional power grid switch cabinet operation parameter data L.

[0254] S3: constructing a power grid switch cabinet operating status identification model, wherein the power grid switch cabinet operating status identification model takes low-dimensional power grid switch cabinet operating parameter data as input and takes the power grid switch cabinet operating status as output.

[0255] The step S3 constructs a power grid switch cabinet operation status identification model, including:

[0256] Constructing a power grid switch cabinet operation state identification model, wherein dynamic latent variable analysis is an implementation method of the power grid switch cabinet operation state identification model, the power grid switch cabinet operation state identification model takes low-dimensional power grid switch cabinet operation parameter data as input and takes power grid switch cabinet operation state as output, wherein the power grid switch cabinet operation state identification model includes an input layer, a latent variable extraction layer, a dynamic representation layer and an identification layer;

[0257] The input layer is used to receive low-dimensional grid switchgear operating parameter data;

[0258] The latent variable extraction layer is used to extract latent variables from the low-dimensional grid switchgear operating parameter data;

[0259] The dynamic representation layer is used to dynamically represent the latent variables and obtain the dynamic latent variables of the low-dimensional power grid switchgear operating parameter data;

[0260] The recognition layer is used to identify the states of dynamic latent variables, where the states include sleep, normal operation, and abnormality.

[0261] S4: Input the low-dimensional grid switch cabinet operating parameter data into the grid switch cabinet operating state recognition model to obtain the grid switch cabinet operating state and perform abnormal monitoring. If an abnormality is detected, an alarm is processed.

[0262] In the step S4, the low-dimensional power grid switch cabinet operating parameter data is input into the power grid switch cabinet operating state recognition model to obtain the power grid switch cabinet operating state and perform abnormal monitoring, including:

[0263] The low-dimensional grid switchgear operating parameter data is input into the grid switchgear operating state recognition model to obtain the grid switchgear operating state, wherein the grid switchgear operating state recognition process is as follows:

[0264] S41: The input layer receives low-dimensional power grid switch cabinet operating parameter data L = (L(1), L(2), ..., L(m), ..., L(M));

[0265] S42: The latent variable extraction layer extracts the latent variable γ in the low-dimensional power grid switchgear operating parameter data L;

[0266] S43: The dynamic characterization layer dynamically characterizes the latent variable γ to obtain the dynamic latent variable Y of the low-dimensional power grid switchgear operating parameter data, wherein the dynamic characterization formula is:

[0267]

[0268] in:

[0269] Q hrepresents the convolution matrix of the hth scale, H represents the maximum scale, and * represents the convolution operator; the convolution steps of convolution matrices of different scales are different;

[0270] S44: The recognition layer performs state recognition on the dynamic latent variable Y, where the state recognition formula is:

[0271]

[0272] in:

[0273] State indicates the operating status of the power grid switch cabinet, where ―1, 0, and 1 correspond to abnormal, dormant, and normal operating states respectively;

[0274] Q θ Represents the mapping matrix corresponding to the state θ, and selects The maximum state is taken as the operating state of the power grid switch cabinet, and an alarm is issued if an abnormality is detected.

[0275] In the step S42, the latent variable extraction layer extracts latent variables in the low-dimensional power grid switch cabinet operating parameter data L, including:

[0276] S421: Generate low-dimensional parameter matrix S:

[0277] S=L T L

[0278] S422: Randomly generate mapping coefficient α 0 , and set the current iteration number of the mapping coefficient to d, the maximum iteration number to D, then the dth iteration result of the mapping coefficient is α d , the initial value of d is 0;

[0279] And calculate the weight coefficient w 0 :

[0280] w 0 =[w 0 (1),w 0 (2),...,w 0 (m),...,w 0 (M)]

[0281]

[0282] in:

[0283] w 0 (m) represents the weight of L(m), ||·|| 2 represents the L2 norm;

[0284] Then the dth iteration result of the weight coefficient is w d ;

[0285] S423: Generate mapping coefficient α d With the weight coefficient w d The corresponding intermediate parameter matrix:

[0286]

[0287] in:

[0288] S 1 (α d ) represents the mapping coefficient α d The corresponding intermediate parameter matrix, S 2 (w d ) represents the weight coefficient w d The corresponding intermediate parameter matrix;

[0289] An operator representing a tensor product computation;

[0290] I represents the identity matrix;

[0291] S424: Iterate the mapping coefficients and weight coefficients:

[0292]

[0293] The iterative error term of the mapping coefficient is calculated, where α d+1 The iterative error term is ρ(α d+1 ):

[0294]

[0295] S425: If the iterative error term is less than the preset threshold, α d+1 as the mapping coefficient, otherwise set d=d+1 and return to step S424;

[0296] S426: Based on the mapping coefficient α d+1 , calculate the low-dimensional mapping matrix S' corresponding to the low-dimensional parameter matrix S:

[0297]

[0298] S427: Perform singular value decomposition on the low-dimensional mapping matrix S′, and use the sequence of M singular values ​​obtained by the decomposition as the latent variable γ = (γ(1), γ(2), ..., γ(m), ..., γ(M)), where γ(m) represents the mth singular value obtained by the decomposition.

[0299] Embodiment 2:

[0300] like Figure 2, which is a functional module diagram of a grid switch operating status online monitoring system provided by an embodiment of the present invention, which can implement the grid switch operating status online monitoring method in Example 1.

[0301] The online monitoring system 100 for the operation status of a power grid switch of the present invention can be installed in an electronic device. According to the functions to be implemented, the online monitoring system for the operation status of a power grid switch can include a data acquisition module 101, a dimensionality reduction processing module 102 and a status monitoring device 103. The module of the present invention can also be referred to as a unit, which refers to a series of computer program segments that can be executed by a processor of an electronic device and can complete fixed functions, which are stored in the memory of the electronic device.

[0302] The data acquisition module 101 is used to collect and pre-process the operating parameter data of the power grid switch cabinet to obtain the pre-processed operating parameter data of the power grid switch cabinet;

[0303] A dimensionality reduction processing module 102 is used to perform dimensionality reduction processing on the pre-processed power grid switch cabinet operating parameter data to obtain low-dimensional power grid switch cabinet operating parameter data;

[0304] The state monitoring device 103 is used to construct a grid switch cabinet operation state recognition model, input low-dimensional grid switch cabinet operation parameter data into the grid switch cabinet operation state recognition model, obtain the grid switch cabinet operation state and perform abnormal monitoring, and perform alarm processing if an abnormality is detected.

[0305] In detail, each module in the online monitoring system 100 for the operation status of a power grid switch in the embodiment of the present invention is used in the same manner as described above. Figure 1 The online monitoring method for the operation status of power grid switches described in the present invention has the same technical means and can produce the same technical effects, so it will not be repeated here.

[0306] Embodiment 3:

[0307] like Figure 3 , which is a schematic diagram of the structure of an electronic device for implementing an online monitoring method for the operation status of a power grid switch provided by an embodiment of the present invention.

[0308] The electronic device 1 may include a processor 10 , a memory 11 , a communication interface 13 and a bus, and may also include a computer program stored in the memory 11 and executable on the processor 10 , such as a program 12 .

[0309] The memory 11 includes at least one type of readable storage medium, and the readable storage medium includes a flash memory, a mobile hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory, etc.), a magnetic memory, a disk, an optical disk, etc. The memory 11 may be an internal storage unit of the electronic device 1 in some embodiments, such as a mobile hard disk of the electronic device 1. The memory 11 may also be an external storage device of the electronic device 1 in other embodiments, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device 1. Further, the memory 11 may also include both an internal storage unit of the electronic device 1 and an external storage device. The memory 11 may be used not only to store application software and various types of data installed in the electronic device 1, such as the code of the program 12, etc., but also to temporarily store data that has been output or is to be output.

[0310] The processor 10 may be composed of an integrated circuit in some embodiments, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips, etc. The processor 10 is the control core (Control Unit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and executes or executes programs or modules stored in the memory 11 (such as a program 12 for realizing online monitoring of the operating status of a power grid switch), and calls data stored in the memory 11, so as to execute various functions of the electronic device 1 and process data.

[0311] The communication interface 13 may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device 1 and other electronic devices, and to achieve connection and communication between internal components of the electronic device.

[0312] The bus may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is configured to realize connection and communication between the memory 11 and at least one processor 10, etc.

[0313] Figure 3 Only an electronic device with components is shown, and those skilled in the art will understand that Figure 3 The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0314] For example, although not shown, the electronic device 1 may also include a power source (such as a battery) for supplying power to each component. Preferably, the power source may be logically connected to the at least one processor 10 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include any components such as one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, etc. The electronic device 1 may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0315] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), or a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode) touch device. The display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device 1 and to display a visual user interface.

[0316] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0317] It should be noted that the serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. And the terms "including", "comprising" or any other variants thereof in this article are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "including a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0318] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0319] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for online monitoring of the operating status of a power grid switch, characterized in that: The method comprises: S1: collecting and preprocessing the operating parameter data of the power grid switch cabinet to obtain the preprocessed operating parameter data of the power grid switch cabinet, wherein the operating parameters of the power grid switch cabinet include current, voltage, power, insulation resistance, power grid frequency, harmonic content, temperature and humidity; S2: Performing dimensionality reduction processing on the pre-processed power grid switch cabinet operating parameter data to obtain low-dimensional power grid switch cabinet operating parameter data; S3: constructing a power grid switch cabinet operation status recognition model, wherein the power grid switch cabinet operation status recognition model takes low-dimensional power grid switch cabinet operation parameter data as input and takes the power grid switch cabinet operation status as output; S4: inputting the low-dimensional power grid switch cabinet operation parameter data into the power grid switch cabinet operation state recognition model, obtaining the power grid switch cabinet operation state and performing abnormal monitoring, and performing alarm processing if abnormality is detected; The process of identifying the operating status of the power grid switch cabinet is as follows: S41: The input layer receives low-dimensional power grid switch cabinet operating parameter data L = (L(1), L(2), ..., L(m), ..., L(M)); S42: The latent variable extraction layer extracts the latent variable γ in the low-dimensional power grid switchgear operating parameter data L; S43: The dynamic characterization layer dynamically characterizes the latent variable γ to obtain the dynamic latent variable Y of the low-dimensional power grid switchgear operating parameter data, wherein the dynamic characterization formula is: in: Q h represents the convolution matrix of the hth scale, H represents the maximum scale, and * represents the convolution operator; the convolution steps of convolution matrices of different scales are different; S44: The recognition layer performs state recognition on the dynamic latent variable Y, where the state recognition formula is: in: State indicates the operating status of the power grid switch cabinet, where ―1, 0, and 1 correspond to abnormal, dormant, and normal operating states respectively; Q θ Represents the mapping matrix corresponding to the state θ, and selects The maximum state is taken as the operating state of the power grid switch cabinet, and an alarm is issued if an abnormality is detected.

2. A method for online monitoring of the operating status of a power grid switch according to claim 1, characterized in that: The operation parameter data of the power grid switch cabinet is collected in step S1, including: Collect the operating parameter data of the power grid switch cabinet, where the operating parameters of the power grid switch cabinet include current, voltage, power, insulation resistance, power grid frequency, harmonic content, temperature and humidity. The collected operating parameter data are: in: x represents the operating parameter data; x n The data sequence representing the nth operating parameter index, N represents the total number of operating parameter indexes, wherein the first to the Nth operating parameter indexes are current, voltage, power, insulation resistance, grid frequency, harmonic content, temperature and humidity; Represents the data sequence x n The data values ​​at M sampling moments, Represents the data sequence x n The data value at the mth sampling moment in , m∈[1,M], M represents the number of discrete data values ​​collected in real time during the real-time operation parameter data collection process; The operating parameter data x is preprocessed to obtain preprocessed power grid switch cabinet operating parameter data X.

3. A method for online monitoring of the operating status of a power grid switch according to claim 2, characterized in that: The preprocessing of the operating parameter data x includes: The operating parameter data x is preprocessed, wherein the preprocessing process is: S11: Convert any data sequence in the operating parameter data x into a matrix form, where the data sequence x n The corresponding matrix is: in: E n Represents the data sequence x n The corresponding matrix, S12: Perform singular value decomposition on the transformed matrix to obtain the left and right singular matrices and the singular value diagonal matrix of the matrix, where the matrix E n The left singular matrix of The right singular matrix is The singular value diagonal matrix is ​​∑ n : in: Represents the matrix E n a singular values ​​of , diag(·) indicates diagonal matrixing; S13: Perform spectral representation on any data sequence in the operating parameter data x to obtain the spectral representation result of the data sequence at a number of transformation points, where Data sequence x n The spectrum representation result at the i-th transformation point is: in: Represents the data sequence x n The spectrum representation result at the i-th transformation point; exp(·) represents an exponential function with a natural constant as the base; j represents the imaginary unit; S14: Perform difference change on the spectrum representation result to obtain a difference sequence of the data sequence at a transformation points, where the data sequence x n The corresponding difference sequence is: in: P n Represents the data sequence x n The corresponding difference sequence is The spectrum shows the result and The energy difference between S15: Construct the period threshold of the data sequence, where the data sequence x n The corresponding cycle threshold is: in: T n Represents the data sequence x n The corresponding cycle threshold; σ n Represents the difference sequence P n The standard deviation of μ n Represents the difference sequence P n The mean of S16: The number of singular values ​​required to generate the data sequence, where the data sequence x n The number of singular values ​​to be retained is num n : in: represents an intermediate variable; S17: Filter the singular value diagonal matrix based on the number of singular values ​​and form a denoising matrix of the data sequence, where the data sequence x n The corresponding denoising matrix is: in: T stands for transpose; E′ n Represents the data sequence x n The corresponding denoising matrix, the data distribution form of the denoising matrix is ​​the same as the matrix E n Consistency; ∑′ n Represents a diagonal matrix ∑ that retains singular values n Middle front num n singular values ​​and set the remaining singular values ​​to 0; S18: According to the inverse operation of converting the data sequence into a matrix, a data sequence of length M is extracted from the denoising matrix, and the extracted data sequence is normalized to obtain a normalized data sequence of N operating parameter indicators, wherein the normalized data sequence of the nth operating parameter indicator is: in: X n Represents the normalized data sequence of the nth operating parameter indicator; Represents the normalized data sequence X n The data value at the mth sampling moment in ; The pre-processed grid switch cabinet operating parameter data X = {X n |n∈[1,N]}.

4. A method for online monitoring of the operating status of a power grid switch according to claim 3, characterized in that: In the step S2, the pre-processed grid switch cabinet operating parameter data is subjected to dimensionality reduction processing, including: The pre-processed power grid switch cabinet operating parameter data X is subjected to dimensionality reduction processing to obtain low-dimensional power grid switch cabinet operating parameter data, wherein the dimensionality reduction processing flow is as follows: S21: Convert the pre-processed grid switch cabinet operating parameter data X into a sequence form: X ′ =(X ′ (1),X ′ (2),...,X ′ (m),...,X ′ (M)) in: X ′ represents the sequence representation of the pre-processed power grid switch cabinet operating parameter data X; X ′ (m) represents the operating parameter data sequence at the mth sampling moment; S22: Calculate the conditional probability that different operating parameter data sequences are similar, where the operating parameter data sequence X ′ (m1) and X ′ (m2) Similar conditional probability is: in: p(X ′ (m1);X ′ (m2)) represents the operating parameter data sequence X ′ (m1) and X ′ (m2) Similar conditional probability, m1,m2∈[1,M],m1≠m2; dis(X ′ (m1),X ′ (m2)) represents the operating parameter data sequence X ′ (m1) and X ′ (m2) distance between; δ(m1) represents the operating parameter data sequence X ′ Standard deviation of (m1); W n Indicates the indicator weight of the nth operating parameter indicator; T n Indicates the data sequence x corresponding to the nth operating parameter indicator n The cycle threshold of S23: Initialize and generate low-dimensional structure data L 0 =(L 0 (1),L 0 (2),...,L 0 (m),...,L 0 (M)), where L 0 (m) is the operating parameter data sequence X ′ (m) low-dimensional representation sequence; Set the low-dimensional structure data L 0 The current number of iterations is t, the maximum number of iterations is Max, and the initial value of t is 0; S24: Calculate the conditional probability of similarity of different low-dimensional representation sequences in the low-dimensional structure data, and combine the conditional probability of similarity of the operating parameter data sequence, iterate the low-dimensional structure data, and obtain the low-dimensional power grid switch cabinet operating parameter data L = (L(1), L(2), ..., L(m), ..., L(M)) corresponding to the pre-processed power grid switch cabinet operating parameter data X, where L(m) is the operating parameter data sequence X ′ (m) The corresponding low-dimensional grid switchgear operating parameter representation sequence.

5. A method for online monitoring of the operating status of a power grid switch according to claim 4, characterized in that: In the step S24, the conditional probabilities of similarity of different low-dimensional representation sequences in the low-dimensional structure data are calculated, and the conditional probabilities of similarity of the operating parameter data sequences are combined to iterate the low-dimensional structure data: S241: Calculate the low-dimensional structure data L obtained at the tth iteration t The conditional probability that any two sets of low-dimensional representation sequences are similar, where L t (m1) and L t (m2) Similar conditional probability is: in: q(L t (m1),L t (m2)) represents L t (m1) and L t (m2) Similar conditional probability; ||·|| represents the L1 norm; S242: Constructing the objective function of low-dimensional structure data optimization: in: F(·) represents the objective function, F(L t ) represents the low-dimensional structure data L t The corresponding objective function value; p(X ′ (m1),X ′ (m2)) represents X ′ (m1),X ′ (m2) similar joint probability between; Low-dimensional structured data L t The smaller the corresponding objective function value is, the lower the dimensional structure data L is. t The more consistent the similarity distribution is with the similarity distribution of the grid switchgear operating parameter data; S243: Taking minimizing the objective function as the optimization goal, the low-dimensional structure data L t Iteration is performed, where the iteration formula is: in: η t represents the iteration weight coefficient; represents the gradient of the objective function; v t represents iterative momentum, β represents momentum weight coefficient; S244: Let t=t+1, and return to step S241 until the maximum number of iterations is reached, and the low-dimensional structure data obtained by the iteration at this time is used as the low-dimensional power grid switch cabinet operation parameter data L.

6. A method for online monitoring of the operating status of a power grid switch according to claim 1, characterized in that: The step S3 constructs a power grid switch cabinet operation status identification model, including: Constructing a power grid switch cabinet operation status recognition model, wherein the power grid switch cabinet operation status recognition model takes low-dimensional power grid switch cabinet operation parameter data as input and takes the power grid switch cabinet operation status as output, wherein the power grid switch cabinet operation status recognition model includes an input layer, a latent variable extraction layer, a dynamic representation layer and a recognition layer; The input layer is used to receive low-dimensional grid switchgear operating parameter data; The latent variable extraction layer is used to extract latent variables from the low-dimensional grid switchgear operating parameter data; The dynamic representation layer is used to dynamically represent the latent variables and obtain the dynamic latent variables of the low-dimensional power grid switchgear operating parameter data; The recognition layer is used to identify the states of dynamic latent variables, where the states include sleep, normal operation, and abnormality.

7. A method for online monitoring of the operating status of a power grid switch according to claim 6, characterized in that: In the step S42, the latent variable extraction layer extracts latent variables in the low-dimensional power grid switch cabinet operating parameter data L, including: S421: Generate low-dimensional parameter matrix S: S=L T L S422: Randomly generate a mapping coefficient α0, and set the current iteration number of the mapping coefficient to d, the maximum iteration number to D, and the dth iteration result of the mapping coefficient to α d , the initial value of d is 0; And calculate the weight coefficient w0: w0=[w0(1),w0(2),...,w0(m),...,w0(M)] in: w0(m) represents the weight of L(m), ||·||2 represents the L2 norm; Then the dth iteration result of the weight coefficient is w d ; S423: Generate mapping coefficient α d With the weight coefficient w d The corresponding intermediate parameter matrix: in: S1(α d ) represents the mapping coefficient α d The corresponding intermediate parameter matrix, S2(w d ) represents the weight coefficient w d The corresponding intermediate parameter matrix; An operator representing a tensor product computation; I represents the identity matrix; S424: Iterate the mapping coefficients and weight coefficients: The iterative error term of the mapping coefficient is calculated, where α d+1 The iterative error term is ρ(α d+1 ): S425: If the iterative error term is less than the preset threshold, α d+1 as the mapping coefficient, otherwise set d=d+1 and return to step S424; S426: Based on the mapping coefficient α d+1 , calculate the low-dimensional mapping matrix S corresponding to the low-dimensional parameter matrix S ′ : S427: Low dimensional mapping matrix S ′ Perform singular value decomposition, and use the sequence of M singular values ​​obtained by the decomposition as the latent variable γ = (γ(1), γ(2), ..., γ(m), ..., γ(M)), where γ(m) represents the mth singular value obtained by the decomposition.

8. An online monitoring system for the operation status of a power grid switch, characterized in that: The system comprises: The data acquisition module is used to collect the operating parameter data of the power grid switch cabinet and perform preprocessing to obtain the preprocessed operating parameter data of the power grid switch cabinet; A dimension reduction processing module is used to perform dimension reduction processing on the pre-processed power grid switch cabinet operating parameter data to obtain low-dimensional power grid switch cabinet operating parameter data; A state monitoring device is used to construct a power grid switch cabinet operating state identification model, input low-dimensional power grid switch cabinet operating parameter data into the power grid switch cabinet operating state identification model, obtain the power grid switch cabinet operating state and perform abnormal monitoring, and if an abnormality is detected, perform alarm processing to realize an online monitoring method for the power grid switch operating state as described in any one of claims 1-7.

Images