Security risk assessment method for power distribution network

Abstract

The invention discloses a safety risk assessment method for a power distribution network, and relates to the technical field of power grid safety early warning, and the method comprises the steps of safety data conversion, reference data presetting and risk data early warning. Realizing data conversion to obtain frequency spectrum information; secondly, performing clustering algorithm processing on the frequency spectrum information, determining a core point, calculating related frequency domain data to generate a frequency domain cluster baseline model, finally acquiring data in real time during operation of the method, dividing a current frequency domain cluster, calculating to obtain a cluster drift evaluation index, and calculating an equipment and power distribution network safety evaluation index by combining periodic safety data such as an equipment failure rate and the like; according to the method, the safety risk of the power distribution network can be effectively evaluated, and stable operation of the power distribution network is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid security early warning, and in particular to a security risk assessment method for a power distribution network. Background Art

[0002] In the modern power system, the distribution network is directly related to the reliability and quality of users' electricity consumption. The operating status of the distribution network is more complex and changeable, and the security risks are also increasing. Therefore, a security risk assessment method for the distribution network is needed.

[0003] Existing technologies rely on fixed security risk threshold settings and are unable to adapt to the dynamic changes in distribution network operating parameters. Faced with the large amount of data with different characteristics generated during the operation of the distribution network, traditional methods cannot effectively distinguish the periodicity and non-stationary characteristics of the data, resulting in inaccurate data processing and analysis, making it difficult to achieve a comprehensive and accurate assessment of distribution network security risks.

[0004] Most existing technologies use simple data statistical analysis, which makes it difficult to explore the potential risks behind the data. They do not take into account the correlation and continuity of the data, do not analyze the data set, cannot effectively distinguish the defects of data characteristics, and cannot provide early warning of the direction of data cluster drift. Therefore, they cannot timely and accurately discover and deal with distribution network safety risks. Summary of the Invention

[0005] In view of the above-mentioned technical deficiencies, the purpose of the present invention is to provide a security risk assessment method for a distribution network.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions: The present invention provides a security risk assessment method for a distribution network, comprising the following steps: Step 1, security data conversion: collecting various types of security-related data of the distribution network, and then collecting change data of various types of security-related data of the distribution network, analyzing the change data of various types of security-related data of the distribution network, and obtaining frequency domain conversion schemes for various types of security-related data; based on the frequency domain conversion schemes for various types of security-related data, converting various types of security-related data of the distribution network to obtain spectrum information of various types of security-related data of the distribution network.

[0007] Step 2: Preset baseline data: Classify the spectrum information of various security-related data types according to the frequency domain data characteristics to obtain data clusters. Analyze each data cluster to generate frequency domain data for each data cluster. Then, based on the frequency domain data of each cluster, generate a baseline model for the frequency domain cluster.

[0008] Step 3. Risk data warning: During the operation of the distribution network, data is collected in real time and clustered to obtain the current frequency domain cluster. Based on the baseline model of the frequency domain cluster, the current frequency domain cluster is identified to obtain cluster drift data. The cluster drift data is analyzed to obtain the cluster drift assessment index. At the same time, periodic distribution network safety data is collected, and the periodic distribution network safety data and cluster drift assessment index are analyzed. Warnings are issued based on the analysis results.

[0009] Preferably, the specific early warning process is as follows: obtaining a preset standard equipment safety assessment index, a benchmark equipment safety assessment index, a standard distribution network safety assessment index and a benchmark distribution network safety assessment index from a database.

[0010] When the distribution network safety assessment index is greater than or equal to the benchmark distribution network safety assessment index and less than the standard distribution network safety assessment index, the equipment whose safety assessment index is greater than the benchmark equipment safety assessment index is recorded as a faulty equipment, the operation of each faulty equipment is stopped, and an early warning is issued.

[0011] When the distribution network safety assessment index is greater than or equal to the standard distribution network safety assessment index, all equipment in the distribution network will be stopped and the staff will be prompted to order the equipment safety assessment index from large to small.

[0012] When the distribution network safety assessment index is less than the benchmark distribution network safety assessment index, if the equipment safety assessment index of a certain device is greater than the standard equipment safety assessment index, the device will be recorded as a risk device, so as to obtain the various risk devices, stop the operation of each risk device, and issue an early warning.

[0013] The beneficial effects of the present invention are: 1. The present invention first collects distribution network safety-related data and its change data, and obtains spectrum information through fast Fourier transform or wavelet transform to realize data conversion; secondly, the spectrum information is processed by a clustering algorithm to determine the core point, and the relevant frequency domain data is calculated to generate a frequency domain cluster baseline model. Finally, when the method is running, real-time data is collected to divide the current frequency domain cluster, and the cluster drift evaluation index is calculated. The equipment and distribution network safety evaluation index is calculated in combination with periodic safety data such as equipment failure rate; according to the preset standard evaluation index, faulty equipment and risky equipment are early warned. This method can effectively evaluate the safety risk of the distribution network and ensure the stable operation of the distribution network.

[0014] 2. By calculating the periodicity assessment index, the periodicity and non-stationary characteristics of power grid security-related data are scientifically distinguished, and targeted processing is performed using fast Fourier transform and wavelet transform respectively to accurately obtain spectral information. This invention can effectively distinguish the defects of data characteristics, and at the same time can more comprehensively and accurately mine the information contained in the data, providing a reliable basis for subsequent evaluation.

[0015] 3. The present invention determines the weight factor by statistical data quantity and generates the frequency domain cluster baseline model by weighted summation. The baseline model fully considers the relationship and dynamic changes between the data. Compared with the traditional model that relies on fixed thresholds or simple statistical analysis, it can better adapt to the complex and changeable operating conditions of the distribution network and improve the scientificity and accuracy of the evaluation.

[0016] 4. During the operation of the distribution network, the present invention collects data in real time and calculates cluster drift data and evaluation indexes, conducts comprehensive analysis in combination with periodic distribution network safety data, and implements graded early warning based on preset standard evaluation indexes. This method can promptly discover safety hazards during the operation of the distribution network. This method can improve the stability and safety of the operation of the distribution network, reduce the probability of failures, and reduce the economic losses caused by failures. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 The figure is a flow chart of the steps for implementing the method of the present invention. DETAILED DESCRIPTION

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0020] according to Figure 1 As shown, the present invention provides a security risk assessment method for a distribution network, comprising the following steps: Step 1, security data conversion: collecting various types of security-related data of the distribution network, and then collecting change data of various types of security-related data of the distribution network, analyzing the change data of various types of security-related data of the distribution network, and obtaining frequency domain conversion schemes for various types of security-related data; based on the frequency domain conversion schemes for various types of security-related data, converting various types of security-related data of the distribution network to obtain spectrum information of various types of security-related data of the distribution network.

[0021] In a specific embodiment, the specific process of collecting various types of safety-related data of the power distribution network is as follows: various types of safety-related data of the power distribution network are collected through various sensors.

[0022] It should be noted that various types of safety-related data include but are not limited to current intensity, temperature, stress and humidity, and the corresponding types of sensors include but are not limited to current sensors, temperature sensors, stress sensors and humidity sensors.

[0023] In a specific embodiment, the change data of various types of safety-related data of the distribution network are collected, and the specific collection process is as follows: the change data of various types of safety-related data of the distribution network include the variance change rate, trend line slope and autocorrelation function change of various types of safety-related data of the distribution network.

[0024] Various types of safety-related data of the distribution network at various time points are collected by various sensors at various time points, and the variance of the various types of safety-related data at various time points of the distribution network is calculated to obtain the variance of the various types of safety-related data at various time points of the distribution network. The difference between the variance of the ending time point of the various types of safety-related data and the variance of the starting time point is subtracted, and the result is divided by the variance of the starting time point to obtain the variance change rate of the various types of safety-related data of the distribution network.

[0025] Various types of safety-related data at various time points in the distribution network are plotted with time as the horizontal coordinate and the corresponding physical quantity as the vertical coordinate. The change curves of various types of safety-related data are then fitted using the least squares method to obtain the trend lines of various types of safety-related data. The slopes of the trend lines of various types of safety-related data are obtained through image recognition technology and recorded as the slopes of the trend lines of various types of safety-related data in the distribution network.

[0026] Substitute various types of safety-related data of the distribution network at various time points into the autocorrelation function formula to obtain the autocorrelation function values ​​of various types of safety-related data at various time points. In this way, a curve showing the change of the autocorrelation function values ​​of various types of safety-related data over time is obtained, which is recorded as the autocorrelation function curve graph of various types of safety-related data. Use image recognition technology to obtain the local peaks and valleys of the autocorrelation function curve graph of various types of safety-related data. Subtract the average value of the local valleys from the average value of the local peaks of the autocorrelation function curve graph of various types of safety-related data, and record the absolute value of the difference as the change in the autocorrelation function of various types of safety-related data of the distribution network.

[0027] It should be noted that the autocorrelation function formula is an existing technology and can be found on the Internet, so it will not be described in detail.

[0028] In a specific embodiment, the change data of various types of safety-related data of the distribution network are analyzed, and the specific analysis process is as follows: the variance change rate, trend line slope and autocorrelation function change of various types of safety-related data of the distribution network are substituted into the periodic evaluation index calculation formula to obtain the periodic evaluation index of various types of safety-related data of the distribution network.

[0029] It should be noted that the calculation formula for the periodic assessment index is:

[0030] Among them, G a is the periodic evaluation index of category a safety-related data of the distribution network, a is the category number of the safety-related data, the value of a is a positive integer, e is the natural logarithm, A 1a 、A 2a and A 3a are the variance change rate, trend line slope and autocorrelation function change of Class A safety-related data of the distribution network, respectively. A′1, A′2 and A′3 are the preset standard variance change rate, standard trend line slope and standard autocorrelation function change, respectively. ε1, ε2 and ε3 are the preset variance change rate weight factor, trend line slope weight factor and autocorrelation function change weight factor, respectively. ε1>0, ε2>0, ε3>0, ε1+ε2+ε3=1.

[0031] The standard parameters A′1, A′2 and A′3 are the variance change rate threshold, trend line slope threshold and autocorrelation function change threshold of normal data. When the variance change rate of a certain type of data is greater than the threshold, it indicates that the change type of this type of data tends to be non-stationary. When the trend line slope of a certain type of data is greater than the threshold, it indicates that the change type of this type of data tends to be non-stationary. When the autocorrelation function change of a certain type of data is greater than the threshold, it indicates that the change type of this type of data tends to be periodic. The specific values ​​are set by the staff, for example, A′1 is 0.7, A′2 is 0.6 and A′3 is 1.2. The weight factors ε1, ε2 and ε3 are obtained by the staff through experiments, and the specific values ​​are set by the staff, for example, ε1 is 0.3, ε2 is 0.3 and ε3 is 0.4.

[0032] If the periodic assessment index of a certain type of safety-related data of the distribution network is greater than or equal to the standard periodic assessment index, it indicates that this type of safety-related data of the distribution network is periodically changing, and is recorded as periodically changing safety-related data; if the periodic assessment index of a certain type of safety-related data of the distribution network is less than the standard periodic assessment index, it indicates that this type of safety-related data of the distribution network is non-stationarily changing, and is recorded as non-stationarily changing safety-related data. In this way, various types of periodically changing safety-related data and various types of non-stationarily changing safety-related data of the distribution network are obtained.

[0033] The frequency domain conversion scheme for various safety-related data is as follows: various periodically changing safety-related data of the distribution network are subjected to fast Fourier transform, and various non-stationarily changing safety-related data of the distribution network are subjected to wavelet transform.

[0034] In a specific embodiment, the various types of safety-related data of the distribution network are converted, and the specific conversion process is as follows: the FFT algorithm is implemented using Python's NumPy library, and the various types of periodically changing safety-related data of the distribution network are input into the NumPy library to obtain the amplitudes of the various types of periodically changing safety-related data of the distribution network at various frequencies, and a spectrum diagram with frequency as the horizontal axis and amplitude as the vertical axis is established to obtain a spectrum diagram of the various types of periodically changing safety-related data, and spectrum information is extracted from the spectrum diagram of the various types of periodically changing safety-related data to obtain spectrum information of the various types of periodically changing safety-related data.

[0035] It should be noted that the FFT algorithm is an existing technology and can be found on the Internet, so it will not be described in detail.

[0036] The DWT algorithm is implemented using Python's PyWavelets library, and various types of non-stationary changing safety-related data of the distribution network are input into the PyWavelets library to obtain the approximate coefficients and detail coefficients of various types of non-stationary changing safety-related data of the distribution network at various scales. The fast Fourier transform algorithm is used to convert the approximate coefficients of various types of non-stationary changing safety-related data of the distribution network at various scales into low-frequency spectrum graphs, and the detail coefficients of various types of non-stationary changing safety-related data of the distribution network at various scales into high-frequency spectrum graphs to obtain high-frequency spectrum graphs and low-frequency spectrum graphs of various types of non-stationary changing safety-related data of the distribution network. The high-frequency spectrum graph and the low-frequency spectrum graph are merged to obtain spectrum graphs of various types of non-stationary changing safety-related data, and the spectrum information is extracted from the spectrum graphs of various types of non-stationary changing safety-related data to obtain the spectrum information of various types of non-stationary changing safety-related data.

[0037] It should be noted that the DWT algorithm is an existing technology and can be found on the Internet, so it will not be described in detail.

[0038] The spectrum information of various types of periodically changing security-related data and the spectrum information of various types of non-stationarily changing security-related data are merged to obtain the spectrum information of various types of security-related data.

[0039] Step 2: Preset baseline data: Classify the spectrum information of various security-related data types according to the frequency domain data characteristics to obtain data clusters. Analyze each data cluster to generate frequency domain data for each data cluster. Then, based on the frequency domain data of each cluster, generate a baseline model for the frequency domain cluster.

[0040] In a specific embodiment, the analysis of each data cluster is performed, and the specific analysis process is as follows: the spectrum information of various security-related data types is input into the DBSCAN clustering algorithm to obtain the distance between each data point and other data points, and a circle is drawn with each data point as the center and a preset distance as the neighborhood radius to obtain the neighborhood of each data point. According to the distance between each data point and other data points, the number of data points in the neighborhood of each data point is statistically obtained. When the number of data points in the neighborhood of a data point is greater than the preset minimum number of points, it indicates that the data point is a core point, and each core point is obtained in this way. The number of data points in the neighborhood of each core point and the neighborhood area are obtained in this way. The number of data points in the neighborhood of each core point is divided by the neighborhood area to obtain the density of each core point. The core points obtained with a density greater than the preset density are recorded as valid core points. The data points of each valid core point are aggregated to obtain various types of security frequency domain data.

[0041] It should be noted that the DBSCAN clustering algorithm is an existing technology and can be found on the Internet, so it will not be described in detail.

[0042] Various types of security frequency domain data are recorded as data clusters, and the peak frequency, low-frequency energy proportion and bandwidth corresponding to each data point in each data cluster are obtained from the spectrum information of various types of security-related data.

[0043] The peak frequency, low-frequency energy proportion and bandwidth corresponding to each data point in each data cluster are respectively calculated by means of the mean, and the average peak frequency, average energy distribution proportion and average bandwidth of each data cluster are obtained, thereby obtaining the frequency domain data of each data cluster. The peak frequency, low-frequency energy proportion and bandwidth corresponding to each data point in each data cluster are respectively calculated by means of the standard deviation, and the peak frequency standard deviation, low-frequency energy proportion standard deviation and bandwidth standard deviation of each data cluster are obtained.

[0044] In a specific embodiment, the baseline model for generating the frequency domain cluster is generated in the following specific generation process: counting the data volume of each data cluster to obtain the number of data points of each data cluster, counting the data volume of various safety-related data of the distribution network to obtain the total data volume, dividing the number of data points of each data cluster by the total data volume to obtain the data volume ratio of each data cluster, recording the data volume ratio of each data cluster as a weight factor, and performing weighted summation on the average peak frequency, average energy distribution ratio, average bandwidth, peak frequency standard deviation, low-frequency energy ratio standard deviation and bandwidth standard deviation of each data cluster to obtain the average peak frequency, average energy distribution ratio, average bandwidth, peak frequency standard deviation, low-frequency energy ratio standard deviation and bandwidth standard deviation of the basic frequency domain cluster.

[0045] The average peak frequency, average energy distribution ratio, average bandwidth, peak frequency standard deviation, low-frequency energy ratio standard deviation and bandwidth standard deviation of the basic frequency domain cluster are filled into the basic framework of the model to obtain the baseline model of the frequency domain cluster.

[0046] Step 3. Risk data warning: During the operation of the distribution network, data is collected in real time and clustered to obtain the current frequency domain cluster. Based on the baseline model of the frequency domain cluster, the current frequency domain cluster is identified to obtain cluster drift data. The cluster drift data is analyzed to obtain the cluster drift assessment index. At the same time, periodic distribution network safety data is collected, and the periodic distribution network safety data and cluster drift assessment index are analyzed. Warnings are issued based on the analysis results.

[0047] In a specific embodiment, the current frequency domain cluster is obtained, and the specific acquisition process is as follows: collect various current safety-related data, cluster the various current safety-related data according to the cluster division process of various safety-related data of the distribution network, obtain the current frequency domain cluster, and obtain the average peak frequency, average energy distribution ratio, average bandwidth, cluster radius and characteristic matrix of the current frequency domain cluster from the current frequency domain cluster.

[0048] In a specific embodiment, the current frequency domain cluster is identified, and the specific identification process is as follows: the cluster drift data includes the center position offset, radius change rate and feature similarity of the current frequency domain cluster.

[0049] The average peak frequency of the current frequency domain cluster, the average energy distribution ratio of the current frequency domain cluster, the average bandwidth of the current frequency domain cluster, the average peak frequency of the basic frequency domain cluster, the average energy distribution ratio of the basic frequency domain cluster, and the average bandwidth of the basic frequency domain cluster are substituted into the center position offset calculation formula to obtain the center position offset of the current frequency domain cluster.

[0050] The calculation formula for the center position offset is the Euclidean distance formula. The Euclidean distance formula is an existing technology and can be specifically queried on the Internet, so it will not be repeated here.

[0051] The radius change rate of the current frequency domain cluster is obtained by subtracting the cluster radius of the basic frequency domain cluster from the cluster radius of the current frequency domain cluster and dividing the result by the cluster radius of the basic frequency domain cluster.

[0052] The cosine similarity calculation is performed on the feature matrix of the current frequency domain cluster and the feature matrix of the basic frequency domain cluster to obtain the feature similarity of the current frequency domain cluster.

[0053] It should be noted that the cluster radius and characteristic matrix of the basic frequency domain cluster can be obtained from the baseline model of the frequency domain cluster. The characteristic matrix is ​​the matrix of the data point set.

[0054] In a specific embodiment, the cluster drift data is analyzed, and the specific analysis process is as follows: the center position offset, radius change rate and feature similarity of the current frequency domain cluster are substituted into the cluster drift evaluation index calculation formula to obtain the cluster drift evaluation index of the current frequency domain cluster.

[0055] It should be noted that the cluster drift evaluation index calculation formula is:

[0056] Among them, α is the cluster drift evaluation index of the current frequency domain cluster, B1, B2 and B3 are the center position offset, radius change rate and feature similarity of the current frequency domain cluster, respectively, B′1, B′2 and B′3 are the preset standard center position offset, standard radius change rate and standard feature similarity, respectively, φ1, φ2 and φ3 are the preset center position offset weight factor, radius change rate weight factor and feature similarity weight factor, respectively, φ1>0, φ2>0, φ3>0, φ1+φ2+φ3=1.

[0057] The setting process of standard parameters B′1, B′2 and B′3 is the same as that of standard parameter A′1, and they are all set by staff, for example, B′1 is 0.6, B′2 is 0.75 and B′3 is 0.74. The setting process of weight factors φ1, φ2 and φ3 is the same as that of weight factor ε1, and they are all set by staff, for example, φ1 is 0.2, φ2 is 0.3 and φ3 is 0.5.

[0058] In a specific embodiment, the periodic distribution network safety data is collected, and the specific collection process is as follows: the periodic distribution network safety data includes the failure rate, line load rate, failure rate change trend index and line load rate change trend index corresponding to each device, the number of failures is recorded when the device fails, and the number of failures of each device within the preset time length is obtained by counting, the number of failures of each device within the preset time length is divided by the preset time length to obtain the failure rate corresponding to each device, the preset time length is divided into time periods, and the failure rate of each time period corresponding to each device is obtained, with the time period as the horizontal axis and the failure rate as the vertical axis, the failure rate of each time period corresponding to each device is fitted into a failure rate change curve graph of each device, the slope of the current time period is obtained from the failure rate change curve graph through image recognition technology, the change trend index corresponding to each slope is obtained from the database, and the failure rate change trend index corresponding to each device is obtained.

[0059] The power of each line collected at each time for a preset time is collected through a power meter, and the actual power of each line is calculated as the average value. The rated load of each line is obtained from the database, and the actual power of each line is divided by the rated load to obtain the load rate of each line. The connection relationship between each device and each line is obtained from the database. If a line is connected to a certain device, it indicates that the line is the line of the device. In this way, the lines of each device are obtained, and then the load rate of each line of each device is obtained. The load rate of each line of each device is averaged to obtain the line load rate of each device. According to the analysis method of the failure rate change trend index, the line load rate of each device is analyzed to obtain the line load rate change trend index corresponding to each device.

[0060] In a specific embodiment, the periodic distribution network safety data and cluster drift assessment index are analyzed, and the specific analysis process is as follows: the failure rate, line load rate, failure rate change trend index and line load rate change trend index corresponding to each device are substituted into the distribution network equipment safety assessment index calculation formula to obtain the equipment safety assessment index of each device, and the cluster drift assessment index of the current frequency domain cluster and the safety assessment index of each device are substituted into the distribution network safety assessment index calculation formula to obtain the distribution network safety assessment index.

[0061] It should be noted that the calculation formula for the distribution network equipment safety assessment index is: Among them, β b is the equipment safety assessment index of equipment b, b is the equipment number, b=1,2......n, n>2, the value of n is the total number of equipment, C 1b 、C 2b 、f 1b and f 2b are the failure rate, line load rate, failure rate change trend index, and line load rate change trend of device b, respectively. C′1, C′2, f1′, and f′2 are the preset standard failure rate, standard line load rate, standard failure rate change trend index, and standard line load rate change trend, respectively. and are the preset fault rate weight factor and line load rate weight factor respectively,

[0062] The setting process of the standard parameters C′1, C′2, f1′ and f′2 is the same as that of the standard parameter A′1. They are all set by the staff. For example, C′1 is 0.63, C′2 is 0.55, f1′ is 0.73 and f′2 is 0.74. The weight factor and The setting process of is the same as that of the weight factor ε1, both of which are set by the staff, for example is 0.4 and is 0.6.

[0063] The calculation formula for the distribution network security assessment index is: Among them, γ is the distribution network security assessment index, β′ and α′ are the preset standard equipment security assessment index and standard cluster drift assessment index, respectively, and δ b is the preset weight factor of device b, δ b >0, η1 and η2 are the weight factors of the device security assessment index and the cluster drift assessment index, respectively. η1>0, η2>0, η1+η2=1.

[0064] The setting process of the standard parameters β′ and α′ is the same as that of the standard parameter A′1, and both are set by the staff. For example, β′ is 0.6 and α′ is 0.7, and the weight factor δ b The setting process is the same as that of the weight factor ε1, and both are set by the staff. For example, when n is 3, b = 1, 2, 3, δ1 is 0.3, δ2 is 0.3, and δ3 is 0.4. The setting process of the weight factors η1 and η2 is the same as that of the weight factor ε1, and both are set by the staff, for example, η1 is 0.3 and η2 is 0.7.

[0065] In a specific embodiment, the early warning is performed in the following specific process: obtaining a preset standard equipment safety assessment index, a benchmark equipment safety assessment index, a standard distribution network safety assessment index and a benchmark distribution network safety assessment index from a database.

[0066] When the distribution network safety assessment index is greater than or equal to the benchmark distribution network safety assessment index and less than the standard distribution network safety assessment index, the equipment whose safety assessment index is greater than the benchmark equipment safety assessment index is recorded as a faulty equipment, the operation of each faulty equipment is stopped, and an early warning is issued.

[0067] When the distribution network safety assessment index is greater than or equal to the standard distribution network safety assessment index, all equipment in the distribution network will be stopped and the staff will be prompted to order the equipment safety assessment index from large to small.

[0068] When the distribution network safety assessment index is less than the benchmark distribution network safety assessment index, if the equipment safety assessment index of a certain device is greater than the standard equipment safety assessment index, the device will be recorded as a risk device, so as to obtain the various risk devices, stop the operation of each risk device, and issue an early warning.

[0069] The database is used to store the change trend index corresponding to each slope, the rated load of each line, the connection relationship between each device and each line, the standard equipment safety assessment index, the benchmark equipment safety assessment index, the standard distribution network safety assessment index and the benchmark distribution network safety assessment index.

[0070] The above content is merely an example and explanation of the concept of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined in this specification, they should all fall within the scope of protection of the present invention.

Claims

1. A security risk assessment method for a distribution network, characterized in that: Includes the following modules: Step 1: Security data conversion: Collect various types of security-related data from the distribution network, and then collect change data of various types of security-related data from the distribution network. Analyze the change data of various types of security-related data from the distribution network to obtain frequency domain conversion schemes for various types of security-related data. Based on the frequency domain conversion schemes for various types of security-related data, convert various types of security-related data from the distribution network to obtain spectrum information of various types of security-related data from the distribution network. Step 2: Preset baseline data: Classify the spectrum information of various security-related data types according to the frequency domain data characteristics to obtain data clusters. Analyze each data cluster to generate frequency domain data for each data cluster. Then, based on the frequency domain data of each cluster, generate a baseline model for the frequency domain cluster. Step 3. Risk data warning: During the operation of the distribution network, data is collected in real time and clustered to obtain the current frequency domain cluster. Based on the baseline model of the frequency domain cluster, the current frequency domain cluster is identified to obtain cluster drift data. The cluster drift data is analyzed to obtain the cluster drift assessment index. At the same time, periodic distribution network safety data is collected, and the periodic distribution network safety data and cluster drift assessment index are analyzed. Warnings are issued based on the analysis results.

2. The method for security risk assessment of a distribution network according to claim 1, characterized in that: The change data of various security-related data of the distribution network are analyzed, and the specific analysis process is as follows: The change data of various types of safety-related data of the distribution network include the variance change rate, trend line slope and autocorrelation function change of various types of safety-related data of the distribution network. The variance change rate, trend line slope and autocorrelation function change of various types of safety-related data of the distribution network are substituted into the periodicity evaluation index calculation formula to obtain the periodicity evaluation index of various types of safety-related data of the distribution network; If the periodicity evaluation index of a certain type of safety-related data of the distribution network is greater than or equal to the standard periodicity evaluation index, it indicates that the safety-related data of the distribution network is periodically changing, and is recorded as periodically changing safety-related data; if the periodicity evaluation index of a certain type of safety-related data of the distribution network is less than the standard periodicity evaluation index, it indicates that the safety-related data of the distribution network is non-stationarily changing, and is recorded as non-stationarily changing safety-related data, thereby obtaining various types of periodically changing safety-related data and various types of non-stationarily changing safety-related data of the distribution network; The frequency domain conversion scheme for various safety-related data is as follows: various periodically changing safety-related data of the distribution network are subjected to fast Fourier transform, and various non-stationarily changing safety-related data of the distribution network are subjected to wavelet transform.

3. The method for security risk assessment of a distribution network according to claim 2, characterized in that: The conversion of various safety-related data of the distribution network is carried out as follows: Using Python's NumPy library to implement the FFT algorithm, various types of periodically changing safety-related data of the distribution network are input into the NumPy library to obtain the amplitudes of the various types of periodically changing safety-related data of the distribution network at various frequencies. A spectrum graph with frequency as the horizontal axis and amplitude as the vertical axis is established to obtain spectrum graphs of the various types of periodically changing safety-related data. Spectral information is extracted from the spectrum graphs of the various types of periodically changing safety-related data to obtain spectrum information of the various types of periodically changing safety-related data. The DWT algorithm is implemented using Python's PyWavelets library, and various types of non-stationary changing safety-related data of the distribution network are input into the PyWavelets library to obtain the approximate coefficients and detail coefficients of the various types of non-stationary changing safety-related data of the distribution network at various scales. The fast Fourier transform algorithm is used to convert the approximate coefficients of the various types of non-stationary changing safety-related data of the distribution network at various scales into low-frequency spectrum graphs, and the detail coefficients of the various types of non-stationary changing safety-related data of the distribution network at various scales into high-frequency spectrum graphs to obtain high-frequency spectrum graphs and low-frequency spectrum graphs of the various types of non-stationary changing safety-related data of the distribution network. The high-frequency spectrum graph and the low-frequency spectrum graph are merged to obtain spectrum graphs of the various types of non-stationary changing safety-related data. Spectral information is extracted from the spectrum graphs of the various types of non-stationary changing safety-related data to obtain spectrum information of the various types of non-stationary changing safety-related data. The spectrum information of various types of periodically changing security-related data and the spectrum information of various types of non-stationarily changing security-related data are merged to obtain the spectrum information of various types of security-related data.

4. The method for security risk assessment of a distribution network according to claim 1, wherein: The specific analysis process of analyzing each data cluster is as follows: The spectrum information of various security-related data types is input into the DBSCAN clustering algorithm to obtain the distance between each data point and other data points. A circle is drawn with each data point as the center and the preset distance as the neighborhood radius to obtain the neighborhood of each data point. According to the distance between each data point and other data points, the number of data points in the neighborhood of each data point is statistically obtained. When the number of data points in the neighborhood of a data point is greater than the preset minimum number of points, it indicates that the data point is a core point. In this way, each core point is obtained, and the number of data points and the neighborhood area in the neighborhood of each core point are obtained. The number of data points in the neighborhood of each core point is divided by the neighborhood area to obtain the density of each core point. The core points with a density greater than the preset density are recorded as valid core points. The data points of each valid core point are aggregated to obtain various types of security frequency domain data. Record various types of safety frequency domain data as data clusters, and obtain the peak frequency, low-frequency energy ratio, and bandwidth corresponding to each data point in each data cluster from the spectrum information of various types of safety-related data; The peak frequency, low-frequency energy proportion and bandwidth corresponding to each data point in each data cluster are respectively calculated by means of the mean, and the average peak frequency, average energy distribution proportion and average bandwidth of each data cluster are obtained, thereby obtaining the frequency domain data of each data cluster. The peak frequency, low-frequency energy proportion and bandwidth corresponding to each data point in each data cluster are respectively calculated by means of the standard deviation, and the peak frequency standard deviation, low-frequency energy proportion standard deviation and bandwidth standard deviation of each data cluster are obtained.

5. The method for security risk assessment of a distribution network according to claim 4, characterized in that: The specific generation process of the baseline model for generating frequency domain clusters is as follows: Count the data volume of each data cluster to obtain the number of data points of each data cluster, count the data volume of various safety-related data of the distribution network to obtain the total data volume, divide the number of data points of each data cluster by the total data volume to obtain the data volume ratio of each data cluster, record the data volume ratio of each data cluster as a weight factor, and perform weighted summation on the average peak frequency, average energy distribution ratio, average bandwidth, peak frequency standard deviation, low-frequency energy ratio standard deviation, and bandwidth standard deviation of each data cluster to obtain the average peak frequency, average energy distribution ratio, average bandwidth, peak frequency standard deviation, low-frequency energy ratio standard deviation, and bandwidth standard deviation of the basic frequency domain cluster; The average peak frequency, average energy distribution ratio, average bandwidth, peak frequency standard deviation, low-frequency energy ratio standard deviation and bandwidth standard deviation of the basic frequency domain cluster are filled into the basic framework of the model to obtain the baseline model of the frequency domain cluster.

6. The method for security risk assessment of a distribution network according to claim 4, characterized in that: The current frequency domain cluster is obtained, and the specific acquisition process is as follows: Collect various current safety-related data, cluster the various current safety-related data according to the clustering process of various safety-related data of the distribution network, obtain the current frequency domain cluster, and obtain the average peak frequency, average energy distribution ratio, average bandwidth, cluster radius and characteristic matrix of the current frequency domain cluster from the current frequency domain cluster.

7. The method for security risk assessment of a distribution network according to claim 6, characterized in that: The specific identification process of the current frequency domain cluster is as follows: Cluster drift data includes the center position offset, radius change rate and feature similarity of the current frequency domain cluster; Substitute the average peak frequency of the current frequency domain cluster, the average energy distribution ratio of the current frequency domain cluster, the average bandwidth of the current frequency domain cluster, the average peak frequency of the basic frequency domain cluster, the average energy distribution ratio of the basic frequency domain cluster, and the average bandwidth of the basic frequency domain cluster into the center position offset calculation formula to obtain the center position offset of the current frequency domain cluster; The radius change rate of the current frequency domain cluster is obtained by subtracting the cluster radius of the basic frequency domain cluster from the cluster radius of the current frequency domain cluster and dividing the result by the cluster radius of the basic frequency domain cluster. The cosine similarity calculation is performed on the feature matrix of the current frequency domain cluster and the feature matrix of the basic frequency domain cluster to obtain the feature similarity of the current frequency domain cluster.

8. The method for security risk assessment of a distribution network according to claim 7, characterized in that: The cluster drift data is analyzed, and the specific analysis process is as follows: The center position offset, radius change rate and feature similarity of the current frequency domain cluster are substituted into the cluster drift evaluation index calculation formula to obtain the cluster drift evaluation index of the current frequency domain cluster.

9. The method for security risk assessment of a distribution network according to claim 1, characterized in that: The periodic distribution network security data and cluster drift assessment index are analyzed, and the specific analysis process is as follows: The periodic distribution network safety data includes the failure rate, line load rate, failure rate change trend index, and line load rate change trend index corresponding to each device. The failure rate, line load rate, failure rate change trend index, and line load rate change trend index corresponding to each device are substituted into the distribution network equipment safety assessment index calculation formula to obtain the equipment safety assessment index of each device. The cluster drift assessment index of the current frequency domain cluster and the safety assessment index of each device are substituted into the distribution network safety assessment index calculation formula to obtain the distribution network safety assessment index.

10. The method for security risk assessment of a distribution network according to claim 1, characterized in that: The specific warning process is as follows: Obtaining a preset standard equipment safety assessment index, a benchmark equipment safety assessment index, a standard distribution network safety assessment index, and a benchmark distribution network safety assessment index from a database; When the distribution network security assessment index is greater than or equal to the benchmark distribution network security assessment index and less than the standard distribution network security assessment index, the equipment whose security assessment index is greater than the benchmark equipment security assessment index is recorded as a faulty equipment, and the operation of each faulty equipment is stopped and an early warning is issued; When the distribution network safety assessment index is greater than or equal to the standard distribution network safety assessment index, all equipment in the distribution network will be stopped and the staff will be prompted to change the equipment in descending order of their safety assessment index. When the distribution network safety assessment index is less than the benchmark distribution network safety assessment index, if the equipment safety assessment index of a certain device is greater than the standard equipment safety assessment index, the device will be recorded as a risk device, so as to obtain the various risk devices, stop the operation of each risk device, and issue an early warning.

Images