CN115905360A - Abnormal data measurement identification method and device based on random construction matrix - Google Patents

Abstract

The invention belongs to the technical field of power systems, and discloses an abnormal data measurement identification method and device based on a random construction matrix, wherein the method comprises the following steps: acquiring electrical index data and non-electrical index data of a target power grid within a preset time length; constructing a random matrix based on the electrical index data and the non-electrical index data; determining a target area in the random matrix, and performing matrix transformation on the target matrix in the target area; respectively performing first energy spectrum analysis and second energy spectrum analysis on the transformed matrix to obtain a first screening index and a second screening index; determining the distribution state of the feature root according to the first screening index and the second screening index; and determining a data exception result based on the distribution state of the characteristic root. The method can quickly and accurately identify and screen the abnormal data in the power grid.

Description

A method and device for abnormal data measurement and identification based on random construction matrix

Technical Field

The present invention relates to the technical field of power systems, and in particular to an abnormal data measurement and identification method and device based on a randomly constructed matrix.

Background Art

Smart Electrical Grid (SEG) is a comprehensive intelligent system that develops in parallel and in coordination with the transmission network and power grids at all levels. Its characteristics are the informatization and automation of each voltage level between transmission, substation and distribution. With the continuous construction of SEG, the scale of the power system continues to expand, and the power system has become a typical big data system. The data obtained by the power grid measurement system and the edge perception data such as urban meteorology are of many types, and then inevitably there are problems such as large data dimensions, wide data sources, anomalies in the data, and difficulty in identifying and screening. How to quickly and accurately use the massive data in the power grid and mine the correlation between meteorological factors and load power consumption behavior is of great significance to assist subsequent power dispatching decisions, improve the power grid's ability to respond to extreme weather, and ensure the reliability of power supply to important power users.

Electricity is the basic energy source for people to carry out production activities. With the continuous development of the economy, users have higher and higher requirements for the reliability of power supply; at the same time, the scale of the distribution network continues to grow, and a large amount of multi-dimensional data has been poured into the measurement system, including not only electrical index data, but also non-electrical index data such as meteorological indicators. Inevitably, there are data anomalies, difficulties in identification and screening in these data.

Therefore, how to quickly and accurately identify and screen the massive data in the power grid for anomalies, provide technical support for mining the correlation between non-electrical factors and load power consumption behavior, and assist in subsequent behavioral decisions such as power dispatch, has become an urgent problem to be solved by technical personnel in this field.

Summary of the invention

The embodiments of the present invention provide an abnormal data measurement and identification method and device based on a randomly constructed matrix, so as to be able to quickly and accurately identify and screen the massive data in the power grid for abnormalities, and provide technical support for mining the correlation between non-electrical factors and load power consumption behavior, as well as assisting in subsequent behavioral decisions such as power dispatch. In order to have a basic understanding of some aspects of the disclosed embodiments, a simple summary is given below. This summary section is not a general review, nor is it intended to identify key/important components or describe the scope of protection of these embodiments. Its sole purpose is to present some concepts in a simple form as a preface to the detailed description that follows.

According to a first aspect of an embodiment of the present invention, a method for identifying abnormal data measurement based on a randomly constructed matrix is provided.

In some embodiments, the method comprises:

Obtain electrical index data and non-electrical index data of the target power grid within a preset time period;

Constructing a random matrix based on the electrical index data and the non-electrical index data;

Determine a target region in the random matrix, and perform matrix transformation on the target matrix in the target region to obtain a transformed matrix;

Performing a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix respectively to obtain a first screening index and a second screening index;

Determining a distribution state of characteristic roots according to the first screening index and the second screening index;

The data abnormality result is determined based on the distribution state of the characteristic root.

In some embodiments, the electrical indicator data includes _Na basic state variables, wherein the basic state variables include at least active load data, node voltage data, and branch current data.

In some embodiments, the non-electrical data includes N _b influencing factor variables, wherein the influencing factor variables include at least sunshine data, temperature data, and humidity data of the target location.

In some embodiments, the random matrix constructed based on the electrical indicator data and the non-electrical indicator data is:

Among them, X represents the matrix element, nj represents the number of nodes, and ti represents the time point.

In some embodiments, determining the target area in the random matrix specifically includes:

A movable window is constructed in the random matrix, and a region in the movable window is used as the target region.

In some embodiments, performing matrix transformation on the target matrix in the target area to obtain a transformed matrix specifically includes:

At the target sampling time, the original matrix is obtained from the database;

Convert the original matrix to a standard non-Hermitian matrix;

Calculating a plurality of singular value equivalent matrices of the standard non-Hermitian matrix;

The singular value equivalent matrices are multiplied to form a matrix to be analyzed.

In some embodiments, each of the singular value equivalent matrices is multiplied to form a matrix to be analyzed, and then the method further includes:

Convert the matrix to be analyzed into a standard matrix, and calculate the covariance matrix of the standard matrix;

The covariance matrix is used as the transformed matrix.

In some embodiments, the standard matrix has a mean of 1 and a variance of 0.

In some embodiments, performing a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix respectively to obtain a first screening index and a second screening index specifically includes:

The first energy spectrum analysis is a single-ring rate analysis, and the first screening index is an average spectrum radius.

In some embodiments, the single ring rate analysis specifically includes:

为非Hermitian特征的随机矩阵，

中的每一个元素为符合独立同分布的随机变量，其元素满足以下关系式：set up

is a random matrix of non-Hermitian characteristics,

Each element in is an independent and identically distributed random variable, and its elements satisfy the following relationship:

μ( _xi )＝0, ^σ2 ( _xi )＝1

In the formula, μ( _xi ) represents the mean value, ^σ2 ( _xi ) represents the variance value;

的维度N和T趋于无穷，且保持c＝N/T不变时，奇异值等价矩阵的特征值的经验谱分布收敛到圆环，其中，c表示矩阵行数和列数之比。when

When the dimensions N and T of the matrix tend to infinity and c = N/T remains unchanged, the empirical spectrum distribution of the eigenvalues of the singular value equivalent matrix converges to a ring, where c represents the ratio of the number of matrix rows to the number of columns.

In some embodiments, in the single-ring rate analysis, the probability density function is:

为矩阵特征值，L是奇异值等价矩阵的累积个数，圆环内半径为(1-c)^L/2，圆环外半径为1。In the formula,

is the matrix eigenvalue, L is the cumulative number of singular value equivalent matrices, the inner radius of the ring is (1-c) ^L/2 , and the outer radius of the ring is 1.

In some embodiments, performing a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix respectively to obtain a first screening index and a second screening index specifically includes:

The second energy spectrum analysis is an M-P law analysis, and the second screening index is an M-P curve.

In some embodiments, the M-P law analysis specifically includes:

为非Hermitian特征的随机矩阵，

中的每一个元素为符合独立同分布的随机变量，其元素满足以下关系式：set up

is a random matrix of non-Hermitian characteristics,

Each element in is an independent and identically distributed random variable, and its elements satisfy the following relationship:

μ(x _i )=0,σ ² (x _i )=constant<∞

The covariance matrix is defined as:

Among them, S is the covariance matrix, N is the number of matrix rows, X is the original matrix after data collection, and T is the mathematical symbol representing the matrix transpose;

After matrix transformation, the energy spectrum distribution of the covariance matrix is:

In the formula, λ _S is the eigenvalue of the matrix, c is the ratio of the row and column dimensions of the matrix, which should be between 0 and 1.

Among them, a is the minimum value of the eigenvalue radius distribution in the circular rate, b is the maximum value of the eigenvalue radius distribution in the circular rate, and d is the mean value of the eigenvalue distribution in the circular rate.

In some embodiments, determining the data abnormality result based on the distribution state of the characteristic root specifically includes:

If the eigenvalues are scattered and the average spectrum radius gradually shrinks from the center of the circle, the data anomaly result indicates that the data is abnormal;

If the eigenvalues are evenly distributed and the value of the average spectral radius is stable, the data anomaly result is that there is no data anomaly.

According to a second aspect of an embodiment of the present invention, an abnormal data measurement and identification device based on a randomly constructed matrix is provided.

In some embodiments, the apparatus comprises:

A data acquisition unit, used to acquire electrical index data and non-electrical index data of a target power grid within a preset time period;

A matrix construction unit, used for constructing a random matrix based on the electrical index data and the non-electrical index data;

A matrix transformation unit, used for determining a target region in the random matrix, and performing a matrix transformation on the target matrix in the target region to obtain a transformed matrix;

A spectrum analysis unit, used for performing a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix to obtain a first screening index and a second screening index;

A distribution determination unit, used to determine the distribution state of characteristic roots according to the first screening index and the second screening index;

A result output unit is used to determine the data abnormality result based on the distribution state of the characteristic root.

According to a third aspect of an embodiment of the present invention, a computer device is provided.

In some embodiments, the computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

The abnormal data measurement and identification method based on randomly constructed matrices provided by the present invention obtains the electrical index data and non-electrical index data of the target power grid within a preset time length, and constructs a random matrix based on the electrical index data and the non-electrical index data; determines the target area in the random matrix, and performs matrix transformation on the target matrix in the target area to obtain a transformed matrix; performs a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix respectively to obtain a first screening index and a second screening index; determines the distribution state of the characteristic root according to the first screening index and the second screening index, and determines the data abnormality result based on the distribution state of the characteristic root. The method provided by the present invention detects and identifies anomalies in large amounts of high-dimensional data and locates them, thereby meeting the rapid and efficient use of data, and can quickly and accurately identify and screen the massive data in the power grid for anomalies, and provide technical support for mining the correlation between non-electrical factors and load power consumption behavior, as well as assisting in subsequent behavioral decisions such as power dispatching.

It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

FIG1 is a flow chart of an abnormal data measurement and identification method based on a randomly constructed matrix according to an exemplary embodiment;

FIG2 is a flow chart of random matrix construction provided by the present invention;

FIG3 is a schematic diagram of modeling a random matrix provided by the present invention;

FIG4 is a schematic diagram of a matrix transformation process provided by the present invention;

FIG5 is a schematic diagram of the structure of an abnormal data measurement and identification device based on a randomly constructed matrix according to an exemplary embodiment;

Fig. 6 is a schematic diagram showing the structure of a computer device according to an exemplary embodiment.

Reference numerals:

501 - data acquisition unit, 502 - matrix construction unit, 503 - matrix transformation unit, 504 - quantity spectrum analysis unit, 505 - distribution determination unit, 506 - result output unit.

DETAILED DESCRIPTION

The following description and accompanying drawings fully illustrate the specific embodiments of this article so that those skilled in the art can practice them. Parts and features of some embodiments may be included in or replace parts and features of other embodiments. The scope of the embodiments of this article includes the entire scope of the claims, as well as all available equivalents of the claims. Herein, the terms "first", "second", etc. are only used to distinguish one element from another, without requiring or implying any actual relationship or order between these elements. In fact, the first element can also be called the second element, and vice versa. Moreover, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that the structure, device or equipment including a series of elements includes not only those elements, but also other elements that are not explicitly listed, or also include elements inherent to such structure, device or equipment. In the absence of more restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the structure, device or equipment including the elements. Each embodiment is described in a progressive manner herein, and each embodiment focuses on the differences from other embodiments, and the same and similar parts between the embodiments can be referred to each other.

The terms "longitudinal", "lateral", "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like used herein to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing this document and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In the description herein, unless otherwise specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense, for example, it can be a mechanical connection or an electrical connection, or it can be the internal connection of two elements, it can be a direct connection, or it can be an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

As used herein, the term "plurality" means two or more than two unless otherwise specified.

In this document, the character "/" indicates that the preceding and following objects are in an "or" relationship. For example, A/B means: A or B.

In this article, the term "and/or" is a description of the association relationship between objects, indicating that three relationships may exist. For example, A and/or B means: A or B, or, A and B.

In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

Please refer to FIG. 1 , which is a flow chart showing a method for abnormal data measurement and identification based on a randomly constructed matrix according to an exemplary embodiment.

In a specific embodiment, the abnormal data measurement and identification method based on the randomly constructed matrix provided by the present invention includes the following steps:

S110: Acquire electrical index data and non-electrical index data of the target power grid within a preset time period. Specifically, the electrical index data includes N _a basic state variables, and the non-electrical data includes N _b influencing factor variables. For example, the basic state variables include at least active load data, node voltage data, and branch current data, and the influencing factor variables include at least meteorological data such as sunshine data, temperature data, and humidity data at the target location.

S120: Construct a random matrix based on the electrical index data and the non-electrical index data. The random matrix is:

Among them, X represents the matrix element, nj represents the number of nodes, and ti represents the time point. When constructing the above random matrix, the data measured by the power system measurement system is exported, and then the meteorological data is exported and spliced.

S130: Determine the target region in the random matrix, and perform matrix transformation on the target matrix in the target region to obtain a transformed matrix. When determining the target region, a movable window may be constructed in the random matrix, and the region in the movable window is used as the target region.

In this embodiment, the technical route of the present invention adopts the random construction matrix theory, as shown in Figure 2. A random matrix is constructed for the collected information, and a matrix transformation is performed, wherein the random construction matrix modeling is shown in Figure 3, and the matrix transformation process is shown in Figure 4. In order to effectively reflect the impact of influencing factors on the power grid state and accurately screen the power grid data and influencing factor data, it is necessary to pay attention to the ratio c1 of the dimension of the influencing factor variable and the dimension of the power grid variable when constructing the random matrix. It should be maintained between 0.5 and 1. If the collected influencing factor variable data is less, it is necessary to copy the collected data until the limited requirement of the dimension ratio is met. When the dimension of the random matrix tends to infinity and the row-column ratio c is fixed, according to the random matrix theory, the empirical spectrum distribution of the eigenvalues will converge to the theoretical characteristics. However, in practical applications, as long as the dimension of the matrix is relatively moderate, such as tens to hundreds, a fairly accurate asymptotic convergence result can also be observed, which is the theoretical basis for applying random matrix theory to power system analysis.

S140: performing a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix respectively to obtain a first screening index and a second screening index;

S150: determining a distribution state of characteristic roots according to the first screening index and the second screening index;

S160: Determine the data anomaly result based on the distribution state of the characteristic root; for example, if the characteristic value distribution is scattered and the value of the average spectrum radius gradually shrinks from the center of the circle, then the data anomaly result is that the data is abnormal; if the characteristic value distribution is uniform and the value of the average spectrum radius is stable, then the data anomaly result is that the data is not abnormal.

In principle, the present invention uses a constructed random matrix as a data processing tool, which has the following advantages:

First, the random construction matrix theory in big data technology is a mathematical tool suitable for statistical analysis of complex systems, and its application angle is mainly focused on energy spectrum analysis. In recent years, the random construction matrix theory has been widely used in power systems, especially in the fields of big data reconstruction, situation awareness, abnormal data analysis, fault detection, etc., and has achieved outstanding results. For example, by comparing the eigenvalue distribution of the data matrix when the power system is in normal and abnormal states, combined with the average spectrum radius value, it is judged whether the power system is in an abnormal state; the location of the disturbance or fault can be achieved through the eigenvalue distribution of the regional slice data matrix and the average spectrum radius. The outstanding advantage of the random construction matrix theory is that it can carry large-scale data fusion and has the ability to carry large amounts of multi-dimensional data. Specifically, the random construction matrix can carry regional node voltage, current, phase angle, active load data and other electrical indicators; it can also carry meteorological non-electrical indicators such as light, humidity, and temperature.

Secondly, the random construction matrix has a high direct utilization rate of data, and the data processing is very fast. There is no need for de-unitization, which is suitable for the background of severe meteorological conditions and meets the requirements of rapid processing in critical situations. Meteorological factors are complex and change rapidly, and meteorological factors are increasingly becoming an important factor affecting the load characteristics of power systems. Accurate judgment of load behavior is the basis for stable operation of power grids. At present, the theory of random construction matrices is widely used in fault location and line selection, but it has no application in basic data recognition. In this regard, the present invention discloses an abnormal data measurement and identification system based on a randomly constructed matrix to detect and identify anomalies of large amounts of high-dimensional data and locate them, thereby meeting the rapid and efficient use of data and ensuring the safe operation of the power system.

In FIG4 , a matrix transformation is performed on the target matrix in the target area to obtain a transformed matrix, which specifically includes the following:

At the target sampling time, the original matrix is obtained from the database;

Convert the original matrix to a standard non-Hermitian matrix;

Calculating a plurality of singular value equivalent matrices of the standard non-Hermitian matrix;

Multiplying the singular value equivalent matrices to form a matrix to be analyzed;

The matrix to be analyzed is converted into a standard matrix, and the covariance matrix of the standard matrix is calculated, wherein the mean of the standard matrix is 1 and the variance is 0.

The covariance matrix is used as the transformed matrix.

In some embodiments, performing a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix respectively to obtain a first screening index and a second screening index specifically includes:

The first energy spectrum analysis is a single-ring rate analysis, and the first screening index is an average spectrum radius.

The single ring rate (Ring Law) analysis specifically includes:

为非Hermitian特征的随机矩阵，

中的每一个元素为符合独立同分布的随机变量，其元素满足以下关系式：set up

is a random matrix of non-Hermitian characteristics,

Each element in is an independent and identically distributed random variable, and its elements satisfy the following relationship:

μ( _xi )＝0, ^σ2 ( _xi )＝1

In the formula, μ( _xi ) represents the mean value, ^σ2 ( _xi ) represents the variance value;

的维度N和T趋于无穷，且保持c＝N/T不变时，奇异值等价矩阵的特征值的经验谱分布收敛到圆环，其中，c表示矩阵行数和列数之比。when

When the dimensions N and T of the matrix tend to infinity and c = N/T remains unchanged, the empirical spectrum distribution of the eigenvalues of the singular value equivalent matrix converges to a ring, where c represents the ratio of the number of matrix rows to the number of columns.

Its probability density function is:

为矩阵特征值，L是奇异值等价矩阵的累积个数，圆环内半径为(1-c)^L/2，圆环外半径为1。In the formula,

is the matrix eigenvalue, L is the cumulative number of singular value equivalent matrices, the inner radius of the ring is (1-c) ^L/2 , and the outer radius of the ring is 1.

In some embodiments, performing a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix respectively to obtain a first screening index and a second screening index specifically includes:

The second energy spectrum analysis is an M-P law analysis, and the second screening index is an M-P curve.

The M-P law (Marchenko-Pasturlaw) analysis specifically includes:

为非Hermitian特征的随机矩阵，

中的每一个元素为符合独立同分布的随机变量，其元素满足以下关系式：set up

is a random matrix of non-Hermitian characteristics,

Each element in is an independent and identically distributed random variable, and its elements satisfy the following relationship:

μ(x _i )=0,σ ² (x _i )=constant<∞

The covariance matrix is defined as:

Among them, S is the covariance matrix, N is the number of matrix rows, X is the original matrix after data collection, and T is the mathematical symbol representing the matrix transpose;

After matrix transformation, the energy spectrum distribution of the covariance matrix is:

In the formula, λ _S is the eigenvalue of the matrix, c is the ratio of the row and column dimensions of the matrix, which should be between 0 and 1.

Among them, a is the minimum value of the eigenvalue radius distribution in the circular rate, b is the maximum value of the eigenvalue radius distribution in the circular rate, and d is the mean value of the eigenvalue distribution in the circular rate.

In a specific usage scenario, please still refer to Figures 2 and 3, construct a movable small window in the formed model matrix, take the data in the small window as the research object, and transform the data according to the formula of the M-P law and the single-loop law to obtain the average spectral radius and the M-P curve. The average spectral radius and the M-P curve obtained in the above steps are two indicators for identifying and screening abnormal data in this embodiment. The distribution of characteristic roots can be seen through these two indicators. If the eigenvalue distribution is scattered and the value of the average spectral radius gradually shrinks from the center of the circle, then the data in the small window is abnormal; if the eigenvalue distribution is uniform and the value of the average spectral radius is relatively stable, then the data in the small window is normal.

When a small window data is found to be abnormal, the present invention further comprises the following steps:

First, return the row and column position of the small window data;

Reduce the number of rows and columns of the window, let the small window traverse the entire matrix according to the moving direction in Figure 2, and repeatedly iterate the data processing of the reduced window;

Continue to reduce the number of rows and columns of the abnormal matrix until it is accurately located.

In the above specific implementation, the abnormal data measurement and identification method based on randomly constructed matrix provided by the present invention obtains the electrical index data and non-electrical index data of the target power grid within a preset time, and constructs a random matrix based on the electrical index data and the non-electrical index data; determines the target area in the random matrix, and performs matrix transformation on the target matrix in the target area to obtain the transformed matrix; performs a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix respectively to obtain a first screening index and a second screening index; determines the distribution state of the characteristic root according to the first screening index and the second screening index, and determines the data abnormality result based on the distribution state of the characteristic root. The method provided by the present invention detects and identifies anomalies of large amounts of high-dimensional data and locates them, thereby meeting the rapid and efficient use of data, and can quickly and accurately identify and screen the massive data in the power grid for anomalies, and provide technical support for mining the correlation between non-electrical factors and load power consumption behavior, as well as assisting in subsequent behavioral decisions such as power dispatching.

According to a second aspect of an embodiment of the present invention, an abnormal data measurement and identification device based on a randomly constructed matrix is provided.

In some embodiments, as shown in FIG5 , the apparatus comprises:

The data acquisition unit 501 is used to acquire electrical index data and non-electrical index data of the target power grid within a preset time period;

A matrix construction unit 502, configured to construct a random matrix based on the electrical index data and the non-electrical index data;

A matrix transformation unit 503 is used to determine a target region in the random matrix, and perform a matrix transformation on the target matrix in the target region to obtain a transformed matrix;

The energy spectrum analysis unit 504 is used to perform a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix to obtain a first screening index and a second screening index;

A distribution determination unit 505, configured to determine a distribution state of characteristic roots according to the first screening index and the second screening index;

The result output unit 506 is used to determine the data abnormality result based on the distribution state of the characteristic root.

In some embodiments, the electrical indicator data includes _Na basic state variables, wherein the basic state variables include at least active load data, node voltage data, and branch current data.

In some embodiments, the non-electrical data includes N _b influencing factor variables, wherein the influencing factor variables include at least sunshine data, temperature data, and humidity data of the target location.

In some embodiments, the random matrix constructed based on the electrical indicator data and the non-electrical indicator data is:

Among them, X represents the matrix element, nj represents the number of nodes, and ti represents the time point.

In some embodiments, determining the target area in the random matrix specifically includes:

A movable window is constructed in the random matrix, and a region in the movable window is used as the target region.

In some embodiments, performing matrix transformation on the target matrix in the target area to obtain a transformed matrix specifically includes:

At the target sampling time, the original matrix is obtained from the database;

Convert the original matrix to a standard non-Hermitian matrix;

Calculating a plurality of singular value equivalent matrices of the standard non-Hermitian matrix;

The singular value equivalent matrices are multiplied to form a matrix to be analyzed.

In some embodiments, each of the singular value equivalent matrices is multiplied to form a matrix to be analyzed, and then the method further includes:

Convert the matrix to be analyzed into a standard matrix, and calculate the covariance matrix of the standard matrix;

The covariance matrix is used as the transformed matrix.

In some embodiments, the standard matrix has a mean of 1 and a variance of 0.

In some embodiments, performing a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix respectively to obtain a first screening index and a second screening index specifically includes:

The first energy spectrum analysis is a single-ring rate analysis, and the first screening index is an average spectrum radius.

In some embodiments, the single ring rate analysis specifically includes:

为非Hermitian特征的随机矩阵，

中的每一个元素为符合独立同分布的随机变量，其元素满足以下关系式：set up

is a random matrix of non-Hermitian characteristics,

Each element in is an independent and identically distributed random variable, and its elements satisfy the following relationship:

μ( _xi )＝0, ^σ2 ( _xi )＝1

In the formula, μ( _xi ) represents the mean value, ^σ2 ( _xi ) represents the variance value;

的维度N和T趋于无穷，且保持c＝N/T不变时，奇异值等价矩阵的特征值的经验谱分布收敛到圆环，其中，c表示矩阵行数和列数之比。when

When the dimensions N and T of the matrix tend to infinity and c = N/T remains unchanged, the empirical spectrum distribution of the eigenvalues of the singular value equivalent matrix converges to a ring, where c represents the ratio of the number of matrix rows to the number of columns.

In some embodiments, in the single-ring rate analysis, the probability density function is:

为矩阵特征值，L是奇异值等价矩阵的累积个数，圆环内半径为(1-c)^L/2，圆环外半径为1。In the formula,

is the matrix eigenvalue, L is the cumulative number of singular value equivalent matrices, the inner radius of the ring is (1-c) ^L/2 , and the outer radius of the ring is 1.

In some embodiments, performing a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix respectively to obtain a first screening index and a second screening index specifically includes:

The second energy spectrum analysis is an M-P law analysis, and the second screening index is an M-P curve.

In some embodiments, the M-P law analysis specifically includes:

为非Hermitian特征的随机矩阵，

中的每一个元素为符合独立同分布的随机变量，其元素满足以下关系式：set up

is a random matrix of non-Hermitian characteristics,

Each element in is an independent and identically distributed random variable, and its elements satisfy the following relationship:

μ(x _i )=0,σ ² (x _i )=constant<∞

The covariance matrix is defined as:

Among them, S is the covariance matrix, N is the number of matrix rows, X is the original matrix after data collection, and T is the mathematical symbol representing the matrix transpose;

After matrix transformation, the energy spectrum distribution of the covariance matrix is:

In the formula, λ _S is the eigenvalue of the matrix, c is the ratio of the row and column dimensions of the matrix, which should be between 0 and 1.

Among them, a is the minimum value of the eigenvalue radius distribution in the circular rate, b is the maximum value of the eigenvalue radius distribution in the circular rate, and d is the mean value of the eigenvalue distribution in the circular rate.

In some embodiments, determining the data abnormality result based on the distribution state of the characteristic root specifically includes:

If the eigenvalues are scattered and the average spectrum radius gradually shrinks from the center of the circle, the data anomaly result indicates that the data is abnormal;

If the eigenvalues are evenly distributed and the value of the average spectral radius is stable, the data anomaly result is that there is no data anomaly.

In the above specific implementation, the abnormal data measurement and identification device based on randomly constructed matrix provided by the present invention obtains the electrical index data and non-electrical index data of the target power grid within a preset time, and constructs a random matrix based on the electrical index data and the non-electrical index data; determines the target area in the random matrix, and performs matrix transformation on the target matrix in the target area to obtain the transformed matrix; performs a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix respectively to obtain a first screening index and a second screening index; determines the distribution state of the characteristic root according to the first screening index and the second screening index, and determines the data abnormality result based on the distribution state of the characteristic root. The device provided by the present invention detects and identifies anomalies of large amounts of high-dimensional data and locates them, thereby meeting the rapid and efficient use of data, and can quickly and accurately identify and screen the massive data in the power grid for anomalies, and provide technical support for mining the correlation between non-electrical factors and load power consumption behavior, as well as assisting in subsequent behavioral decisions such as power dispatching.

In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be shown in FIG6. The computer device includes a processor, a memory, and a network interface connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store static information and dynamic information data. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the steps in the above method embodiment are implemented.

Those skilled in the art will understand that the structure shown in FIG. 6 is merely a block diagram of a partial structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

The present invention is not limited to the structures which have been described above and shown in the drawings, and various modifications and changes may be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims (16)

1. A method for measuring and identifying abnormal data based on a randomly constructed matrix, characterized in that the method comprises: Obtain electrical index data and non-electrical index data of the target power grid within a preset time period; Constructing a random matrix based on the electrical index data and the non-electrical index data; Determine a target region in the random matrix, and perform matrix transformation on the target matrix in the target region to obtain a transformed matrix; Performing a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix respectively to obtain a first screening index and a second screening index; Determining a distribution state of characteristic roots according to the first screening index and the second screening index; The data abnormality result is determined based on the distribution state of the characteristic root. 2. According to the abnormal data measurement and identification method based on randomly constructed matrix according to claim 1, it is characterized in that the electrical index data includes _Na basic state variables, wherein the basic state variables include at least active load data, node voltage data and branch current data. 3. According to the abnormal data measurement and identification method based on randomly constructed matrix according to claim 1, it is characterized in that the non-electrical data includes N _b influencing factor variables, wherein the influencing factor variables at least include sunshine data, temperature data and humidity data of the target location. 4. The abnormal data measurement and identification method based on randomly constructed matrix according to claim 1 is characterized in that the random matrix constructed based on the electrical index data and the non-electrical index data is:

Among them, X represents the matrix element, nj represents the number of nodes, and ti represents the time point. 5. The abnormal data measurement and identification method based on the random construction matrix according to claim 1 is characterized in that determining the target area in the random matrix specifically includes: A movable window is constructed in the random matrix, and a region in the movable window is used as the target region. 6. The abnormal data measurement and identification method based on randomly constructed matrix according to claim 1 is characterized in that performing matrix transformation on the target matrix in the target area to obtain the transformed matrix specifically comprises: At the target sampling time, the original matrix is obtained from the database; Convert the original matrix to a standard non-Hermitian matrix; Calculating a plurality of singular value equivalent matrices of the standard non-Hermitian matrix; The singular value equivalent matrices are multiplied to form a matrix to be analyzed. 7. The method for measuring and identifying abnormal data based on randomly constructed matrices according to claim 6, characterized in that each of the singular value equivalent matrices is multiplied to form a matrix to be analyzed, and then further comprising: Convert the matrix to be analyzed into a standard matrix, and calculate the covariance matrix of the standard matrix; The covariance matrix is used as the transformed matrix. 8. The abnormal data measurement and identification method based on randomly constructed matrix according to claim 7 is characterized in that the mean of the standard matrix is 1 and the variance is 0. 9. The abnormal data measurement and identification method based on a randomly constructed matrix according to any one of claims 1 to 8, characterized in that the transformed matrix is subjected to a first energy spectrum analysis and a second energy spectrum analysis respectively to obtain a first screening index and a second screening index, specifically comprising: The first energy spectrum analysis is a single-ring rate analysis, and the first screening index is an average spectrum radius. 10. The abnormal data measurement and identification method based on randomly constructed matrix according to claim 9, characterized in that the single-loop rate analysis specifically includes: set up

is a random matrix of non-Hermitian characteristics,

Each element in is an independent and identically distributed random variable, and its elements satisfy the following relationship: μ( _xi )＝0, ^σ2 ( _xi )＝1 In the formula, μ( _xi ) represents the mean value, ^σ2 ( _xi ) represents the variance value; when

When the dimensions N and T of the matrix tend to infinity and c = N/T remains unchanged, the empirical spectrum distribution of the eigenvalues of the singular value equivalent matrix converges to a ring, where c represents the ratio of the number of matrix rows to the number of columns. 11. The abnormal data measurement and identification method based on random construction matrix according to claim 10 is characterized in that, in the single-loop rate analysis, its probability density function is:

In the formula,

is the matrix eigenvalue, L is the cumulative number of singular value equivalent matrices, the inner radius of the ring is (1-c) ^L/2 , and the outer radius of the ring is 1. 12. The abnormal data measurement and identification method based on a randomly constructed matrix according to any one of claims 1 to 8, characterized in that the transformed matrix is subjected to a first energy spectrum analysis and a second energy spectrum analysis respectively to obtain a first screening index and a second screening index, specifically comprising: The second energy spectrum analysis is an M-P law analysis, and the second screening index is an M-P curve. 13. The abnormal data measurement and identification method based on randomly constructed matrix according to claim 12 is characterized in that the M-P law analysis specifically includes: set up

is a random matrix of non-Hermitian characteristics,

Each element in is an independent and identically distributed random variable, and its elements satisfy the following relationship: μ(x _i )=0,σ ² (x _i )=constant<∞ The covariance matrix is defined as:

Among them, S is the covariance matrix, N is the number of matrix rows, X is the original matrix after data collection, and T is the mathematical symbol representing the matrix transpose; After matrix transformation, the energy spectrum distribution of the covariance matrix is:

In the formula, λ _S is the eigenvalue of the matrix, c is the ratio of the row and column dimensions of the matrix, which should be between 0 and 1.

Among them, a is the minimum value of the eigenvalue radius distribution in the circular rate, b is the maximum value of the eigenvalue radius distribution in the circular rate, and d is the mean value of the eigenvalue distribution in the circular rate. 14. The abnormal data measurement and identification method based on a randomly constructed matrix according to any one of claims 1 to 8, characterized in that determining the data abnormality result based on the distribution state of the characteristic root specifically comprises: If the eigenvalues are scattered and the average spectrum radius gradually shrinks from the center of the circle, the data anomaly result indicates that the data is abnormal; If the eigenvalues are evenly distributed and the value of the average spectral radius is stable, the data anomaly result is that there is no data anomaly. 15. An abnormal data measurement and identification device based on a randomly constructed matrix, characterized in that the device comprises: A data acquisition unit, used to acquire electrical index data and non-electrical index data of a target power grid within a preset time period; A matrix construction unit, used for constructing a random matrix based on the electrical index data and the non-electrical index data; A matrix transformation unit, used for determining a target region in the random matrix, and performing a matrix transformation on the target matrix in the target region to obtain a transformed matrix; A spectrum analysis unit, used for performing a first energy spectrum analysis and a second energy spectrum analysis on the transformed matrix to obtain a first screening index and a second screening index; A distribution determination unit, used to determine the distribution state of characteristic roots according to the first screening index and the second screening index; A result output unit is used to determine the data abnormality result based on the distribution state of the characteristic root. 16. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein the processor implements the steps of any one of the methods of claims 1 to 14 when executing the computer program.

Images