A method for analyzing power grid operation data based on deep learning

Abstract

The present invention discloses a method for analyzing power grid operation data based on deep learning. By setting up a first comparison module, a first database, a detection device, a reminder module, a first processing module, and a first reading module, it can detect the operation data of the power grid. If there are problems with the operation data of the power grid, a message will be sent to the person in charge of the power grid to give an early warning of the discovered power grid problems in a timely manner. By setting up a second comparison module, a second database, a second processing module, and a first reading module, it can diagnose in advance the problems that occur in the power grid and arrange maintenance personnel to conduct inspection and maintenance in advance, without the need to adopt the traditional method, that is, arranging personnel to check the situation first and then arranging personnel to carry out maintenance, which wastes a lot of time and energy.

Description

Technical Field

[0001] The present invention relates to computer technology, in particular to a method for analyzing power grid operation data based on deep learning. Background Art

[0002] The overall power grid consists of various voltage substations and power transmission and distribution lines in the power system, which is called the power grid for short. When the power grid is operating, some electrical components may be damaged. If the damage of electrical components cannot be detected and repaired in time, it will not only cause losses to social property, but also seriously affect people's lives and work. Summary of the Invention

[0003] Object of the Invention: The object of the present invention is to provide a method for analyzing power grid operation data based on deep learning, so as to detect the operation data of the power grid and give early warnings for the discovered power grid problems in time.

[0004] Technical Solution: A method for analyzing power grid operation data based on deep learning according to the present invention includes the following steps:

[0005] S1. Connect the detection device to the power grid. The detection device detects the operation status of the power grid and sends the detected data to the first processing module.

[0006] S2. The first processing module activates the first reading module. The first reading module reads the normal operation data of the power grid in the first database and sends the data to the first processing module.

[0007] S3. The first processing module activates the first comparison module. The first comparison module compares and diagnoses the detected data (current and voltage) with the normal operation data of the power grid to obtain whether the operation data of the power grid is normal, and sends the comparison and diagnosis result to the first processing module.

[0008] S4. When the detected data (current and voltage) of the detection device is inconsistent with the normal operation data of the power grid, the first processing module sends the detected data (current and voltage) to the first transmission module.

[0009] S5. The first transmission module transmits the detected data (current and voltage) to the second transmission module, and the second transmission module sends the detected data (current and voltage) to the second processing module.

[0010] S51. is a continuous testing function, λ is an eigenvalue, N is the spatial dimension, is the phase of current and voltage, i and n are vectors.

[0011] S52. The law of large numbers of LES: converges in probability to: In this equation, p(λ) is the probability density function (PDF) of the matrix eigenvalues.

[0012] S53, LES central limit: Given a non-Hermitian The rectangular matrix x, whose elements Xij satisfy the standard normal independent and identical distribution (iid); M is the covariance sinθ matrix of x. Let the test function satisfy continuous and then N, T→∞ and c=N / T≤1, its value distribution constructed according to formula (2) converges to a Gaussian variable with mean 0 and variance as follows: In, yes The fourth-order cumulant of the element Ψ; θ, θ1, θ2 are the independent variables of integration, K4=E(X 4 )-3(E(X 4 ) is the expected value.

[0013] The expected mean of S54 and LES is calculated according to the law of large numbers formula (2), and the expected variance of LES is calculated according to the central limit theorem formula (3); the expected value and variance deviation are Δ mean = O(N-1), Δ variance = O(N-2).

[0014] S55. By applying the high-dimensional statistical feature construction process of the power grid fault to the real-time data of the power grid, the obtained characteristic value is compared with the expectation and variance to determine whether it exceeds the range set by the expectation and variance, thereby obtaining a preliminary fault diagnosis Rψ.

[0015] S6. The second processing module starts the second reading module, and the second reading module reads a variety of abnormal power grid operation data in the second database and sends the data to the second processing module.

[0016] S7. The second processing module starts the second comparison module, which compares and diagnoses the detection data (current and voltage) with the abnormal operation data of the power grid, obtains the problems of the power grid operation data, and sends the comparison diagnosis results to the second processing module.

[0017] S8. The second processing module sends the problems found in the grid operation data obtained through comparison and diagnosis to the display module.

[0018] Before the step S1 is started, the normal operation data of the power grid and the various abnormal operation data of the power grid are stored in the first database and the second database respectively.

[0019] A computer storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned power grid operation data analysis method based on deep learning.

[0020] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for analyzing power grid operation data based on deep learning is implemented.

[0021] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0022] 1. The present invention is capable of detecting the operation data of the power grid by providing a first comparison module, a first database, a detection device, a reminder module, a first processing module, and a first reading module. If there is a problem with the operation data of the power grid, a message will be sent to the person in charge of the power grid to promptly issue an early warning of the discovered power grid problem;

[0023] 2. The present invention can diagnose problems in the power grid in advance and arrange maintenance personnel to perform inspections and maintenance in advance by setting up a second comparison module, a second database, a second processing module and a first reading module. There is no need to adopt the traditional method of arranging personnel to check the situation first and then arranging personnel to perform maintenance, which wastes a lot of time and energy. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is an operation flow chart of the present invention. DETAILED DESCRIPTION

[0025] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0026] like Figure 1 As shown, a power grid operation data analysis method based on deep learning includes the following steps:

[0027] S1. Connect the detection device to the power grid. The detection device detects the operation status of the power grid and sends the detected data to the first processing module.

[0028] S2. The first processing module starts the first reading module. The first reading module reads the normal operation data of the power grid in the first database and sends the data to the first processing module.

[0029] S3. The first processing module starts the first comparison module, which compares and diagnoses the detection data (current and voltage) with the normal operation data of the power grid to determine whether the power grid operation data is normal, and sends the comparison diagnosis result to the first processing module.

[0030] S4. When the detection data (current and voltage) of the detection device is inconsistent with the normal operation data of the power grid, the first processing module sends the detection data (current and voltage) to the first transmission module.

[0031] S5. The first transmission module transmits the detection data (current and voltage) to the second transmission module, and the second transmission module sends the detection data (current and voltage) to the second processing module.

[0032] S51. is the continuous testing function, λ is the eigenvalue, N is the spatial dimension, are the phases of current and voltage, and i and n are vectors.

[0033] S52. The law of large numbers of LES: converges in probability to: where p(λ) is the probability density function (PDF) of the matrix eigenvalues.

[0034] S53. The central limit of LES: Given a non-Hermitian rectangular matrix x, whose elements Xij satisfy the standard normal independent and identically distributed (i.i.d.); M is the covariance sinθ matrix of x. Let the testing function be continuous and then when N, T → ∞ and c = N / T ≤ 1, the value distribution constructed according to Equation (2) converges to a Gaussian variable with a mean of 0 and a variance as follows: where is the fourth-order cumulant of element Ψ; θ, θ1, θ2 are the integration independent variables, K4 = E(X 4 ) - 3(E(X 4 ) is the expected value.

[0035] S54. The expected mean of LES is calculated according to the law of large numbers formula (2), and the expected variance of LES is calculated according to the central limit theorem formula (3); the deviation between the expected value and the variance is Δ mean = O(N - 1), Δ variance = O(N - 2);

[0036] S55. By applying the construction process of the high-dimensional statistical characteristics of the power grid fault to the real-time power grid data, that is, comparing the obtained eigenvalues with the expected values and variances to determine whether they exceed the set range of the expected values and variances, the preliminary fault diagnosis Rψ can be obtained.

[0037] S6. The second processing module starts the second reading module. The second reading module reads various abnormal operation data of the power grid in the second database and sends the data to the second processing module.

[0038] S7. The second processing module starts the second comparison module. The second comparison module compares and diagnoses the detection data (current and voltage) with the abnormal operation data of the power grid, obtains the problems existing in the power grid operation data, and sends the comparison and diagnosis results to the second processing module.

[0039] S8. The second processing module sends the problems in the grid operation data obtained by the comparison diagnosis to the display module.

[0040] Before the start of step S1, the normal grid operation data and various abnormal grid operation data are respectively stored in the first database and the second database.

[0041] A computer storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the above-mentioned method for analyzing grid operation data based on deep learning.

[0042] A computer device, comprising a storage, a processor and a computer program stored on the storage and executable on the processor, and when the processor executes the computer program, it implements the above-mentioned method for analyzing grid operation data based on deep learning.

Claims

1. A method for analyzing power grid operation data based on deep learning, characterized in that, Including the following steps: S1. Connect the detection device to the power grid. The detection device detects the operation status of the power grid and sends the detected data to the first processing module; S2. The first processing module activates the first reading module. The first reading module reads the normal operation data of the power grid in the first database and sends the data to the first processing module; S3. The first processing module activates the first comparison module. The first comparison module compares and diagnoses the detected data with the normal operation data of the power grid to obtain whether the operation data of the power grid is normal, and sends the comparison and diagnosis result to the first processing module; S4. When there is an inconsistency between the detected data of the detection device and the normal operation data of the power grid, the first processing module sends the detected data to the first transmission module; S5. The first transmission module transmits the detected data to the second transmission module, and the second transmission module sends the detected data to the second processing module; S51、 is a continuous test function, λ is the eigenvalue, N is the spatial dimension, are the phases of current and voltage, and i and n are vectors; S52. Law of Large Numbers of LES: Converges in probability to: where p(λ) is the probability density function of the matrix eigenvalues; S53. Central Limit of LES: Given a non-Hermitian rectangular matrix x with elements Xij that are standard normal independent and identically distributed; let the covariance matrix of x be M = sinθ. Let the test function be continuous and then, as N, T → ∞ and c = N / T ≤ 1, the value distribution of the one constructed according to Equation (2) converges to a Gaussian variable with a mean of 0 and a variance as follows: In, is the fourth cumulant of the element Ψ; θ, θ1, θ2 are the integration independent variables, K4 = E(X 4 ) - 3(E(X 4 ) is the expected value; S54. The expected mean of LES is calculated according to the law of large numbers formula (2), and the expected variance of LES is calculated according to the central limit theorem formula (3); the deviation of the expected value and variance is Δ mean = O(N - 1), Δ variance = O(N - 2); S55. By applying the high-dimensional statistical feature construction process of power grid faults to the real-time data of the power grid, that is, comparing the obtained eigenvalue with the expectation and variance, it is judged whether it exceeds the range set by the expectation and variance to obtain the preliminary fault diagnosis Rψ; S6. The second processing module activates the second reading module. The second reading module reads various abnormal operation data of the power grid in the second database and sends the data to the second processing module; S7. The second processing module activates the second comparison module. The second comparison module compares and diagnoses the detected data with the abnormal operation data of the power grid to obtain the problems existing in the operation data of the power grid, and sends the comparison and diagnosis result to the second processing module; S8. The second processing module sends the problems existing in the operation data of the power grid obtained by the comparison and diagnosis to the display module.

2. The method for analyzing power grid operation data based on deep learning according to claim 1, wherein Before the start of step S1, the normal operation data of the power grid and various abnormal operation data of the power grid are respectively stored in the first database and the second database.

3. A computer storage medium, on which a computer program is stored, characterized in that, When the computer program is executed by a processor, it implements a method for analyzing power grid operation data based on deep learning as described in any one of claims 1-2.

4. A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements a method for analyzing power grid operation data based on deep learning as described in any one of claims 1-2.

Images