Power transmission network fault hybrid diagnosis method and system based on intelligent AI

Abstract

The invention belongs to the technical field of intelligent power grids, and discloses a power transmission network fault hybrid diagnosis method and system based on intelligent AI. Comprising the following steps: collecting # imgabs0 # group fault electrical quantity data; standardization processing is carried out on each group of fault electrical quantity data, and a fault data set is constructed; extracting a fault feature vector from the fault data set; acquiring protection detection data, fusing the fault feature vector and the protection detection data, and positioning a fault occurrence position by adopting a hybrid diagnosis strategy; power grid topological data are obtained, geographic coordinate correction is carried out on the fault occurrence position in combination with the power grid topological data and a geographic information system, and the fault geographic position is obtained; calculating a fault influence range and a fault severity degree according to the fault geographic position and the fault feature vector, and generating a fault diagnosis report; the fault diagnosis robustness and response speed can be effectively improved, the fault occurrence position can be accurately positioned, the fault recovery time can be shortened, and the safety and stability of the power transmission network can be enhanced.

Description

Technical Field

[0001] The present invention relates to the field of smart grid technology, and more specifically, to a hybrid diagnosis method and system for power transmission network faults based on intelligent AI. Background Art

[0002] With the continuous growth of global electricity demand and the rapid development of smart grids, the scale and complexity of transmission networks are also increasing significantly; however, transmission networks often face challenges such as frequent failures, difficult positioning and slow recovery during operation; power system failures will not only cause large-scale power outages, but may also have a serious impact on the safe and stable operation of the power grid; therefore, fast and accurate diagnosis of transmission network faults has become a key task in power system management; in traditional transmission network fault diagnosis, a single data source or a single algorithm is usually used for positioning, which makes it easy to be affected by errors in complex fault scenarios and difficult to achieve high-precision, fast and robust diagnosis.

[0003] With the continuous development of intelligent technology, fault diagnosis methods based on artificial intelligence have gradually become a research hotspot; for example, the patent with publication number CN113030647A discloses an active early warning system for three-span section faults in a transmission network; it includes an accident signal acquisition module, a fault distance calculation module, a transmission network fault diagnosis module, a transmission line ledger maintenance module, a "three-span" section maintenance module, a power grid geographic information display module and an early warning notification module; this invention can realize the visual management of the transmission network and the "three-span" data based on the geographic information platform, can summarize accident signals to realize fault location diagnosis and active early warning of faults in the "three-span" section, can promptly initiate notifications to on-site operation and maintenance personnel, and improve the response speed to faults in the "three-span" section of the transmission network.

[0004] However, although the above-mentioned technology can realize the fault location diagnosis of the transmission network, it only relies on the accident tripping signal for fault distance measurement. The data source is single, and there is a lack of in-depth analysis and extraction of fault characteristics. It is difficult to accurately characterize the fault characteristics, resulting in insufficient ability to identify complex fault modes. Moreover, it mainly relies on traditional ranging calculation methods. In the case of complex power grid structures or multi-section faults, it is easy to be affected by noise interference or line parameter errors, resulting in reduced diagnostic accuracy and insufficient adaptability to different fault types.

[0005] In view of this, the present invention proposes a hybrid diagnosis method and system for power transmission network faults based on intelligent AI to solve the above problems. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned purpose, the present invention provides the following technical solution: a hybrid diagnosis method for power transmission network faults based on intelligent AI, comprising: collection Group fault electrical quantity data, is an integer greater than 1; Each set of fault electrical quantity data is standardized to construct a fault data set; By using wavelet transform and deep learning technology, the fault feature vector is extracted from the fault data set. The fault feature vector includes fault waveform features, fault transient features and fault spectrum features. Obtain protection detection data, fuse fault feature vectors and protection detection data, and use a hybrid diagnosis strategy to quickly locate the fault location; Obtain the grid topology data, combine the grid topology data with the geographic information system, perform geographic coordinate correction on the fault location, and obtain the fault geographical location; According to the fault geographical location and fault feature vector, the fault impact range and fault severity are calculated, and a fault diagnosis report is generated.

[0007] Furthermore, the fault electrical quantity data includes voltage waveform data and current waveform data; The method for constructing a fault data set comprises: De-noising is performed on each group of fault electrical quantities to obtain preliminary processing data, which includes preliminary voltage processing data and preliminary current processing data; Count the number of data in each group of preliminary voltage processing data respectively and mark them as data number; compare each data number respectively, mark the data number with the largest value as the maximum number, and mark the preliminary voltage processing data corresponding to the maximum number as voltage standard data; mark the preliminary current processing data with the same fault recording device as the voltage standard data as the current standard data; mark the preliminary voltage processing data not marked as voltage standard data as voltage data to be aligned, and mark the preliminary current processing data not marked as current standard data as current data to be aligned; Each group of voltage data to be aligned is time-aligned with the voltage standard data to obtain voltage alignment data; each group of current data to be aligned is time-aligned with the current standard data to obtain current alignment data; a fault data set is constructed according to each group of voltage alignment data, current alignment data, voltage standard data and current standard data; wherein, the method of time-aligning the voltage data to be aligned with the voltage standard data is consistent with the method of time-aligning the current data to be aligned with the current standard data.

[0008] Furthermore, the method for time-aligning the voltage data to be aligned with the voltage standard data includes: According to the time sequence, the voltages in the voltage data to be aligned and the voltage standard data are obtained respectively, and the corresponding voltage sequences are constructed respectively; the Euclidean distance between every two voltages in the two voltage sequences is calculated respectively, and marked as voltage distance; a distance matrix is ​​constructed according to the voltage distance, and the elements in the distance matrix correspond to the voltage distance one by one; according to the distance matrix, a cumulative distance matrix is ​​constructed, and the size of the cumulative distance matrix is ​​consistent with the size of the distance matrix; the elements in the first row and the first column of the cumulative distance matrix are marked as starting elements, the elements in the first row of the cumulative distance matrix are marked as head elements, the elements in the first column of the cumulative distance matrix are marked as index elements, and the elements in the cumulative distance matrix that are not marked as starting elements, head elements and index elements are all marked as data elements; The starting point element in the cumulative distance matrix is ​​equal to the starting point element in the distance matrix. Each head element in the cumulative distance matrix is ​​the previous head element plus the corresponding element in the distance matrix. Each index element in the cumulative distance matrix is ​​the previous index element plus the corresponding element in the distance matrix. Each data element in the cumulative distance matrix is ​​the minimum value of the corresponding adjacent elements plus the corresponding element in the distance matrix. The adjacent elements corresponding to the data element include the element located to the left of the data element, the element located above the data element, and the element located to the upper left of the data element in the cumulative distance matrix. Starting from the element corresponding to the lower right corner of the cumulative distance matrix, backtrack to the starting element of the cumulative distance matrix to obtain the optimal path; the backtracking process is: mark the element backtracked to as the current element, select the minimum value among the adjacent elements corresponding to the current element in the cumulative distance matrix as the backtracking element, and update the current element as the backtracking element; according to the optimal path, establish a voltage matching relationship, which is the corresponding relationship between the voltage in the voltage data to be aligned and the voltage in the voltage standard data; according to the voltage matching relationship, interpolate the voltage data to be aligned; the method for interpolating the voltage data to be aligned is: for a voltage in the data to be aligned in the optimal path, the voltage in the voltage standard data is corresponding to the voltage in the voltage standard data. When a voltage is generated, the corresponding voltage is marked as the interpolation voltage, and a linear interpolation method is used at the interpolation voltage to generate A new voltage, is an integer greater than 1.

[0009] Further, the fault waveform characteristics include amplitude characteristics and waveform change rate; the fault transient characteristics include transient amplitude and transient duration; the fault spectrum characteristics include spectrum amplitude and frequency ratio; The voltage waveform data after standardization are all marked as voltage standard waveforms, and the current waveform data are all marked as current standard waveforms; the amplitude characteristics include the maximum amplitude and minimum amplitude of the voltage standard waveform, and the maximum amplitude and minimum amplitude of the current standard waveform; the waveform change rate includes the maximum instantaneous slope of the voltage standard waveform and the maximum instantaneous slope of the current standard waveform; the transient amplitude includes the maximum instantaneous deviation value of the voltage standard waveform and the maximum instantaneous deviation value of the current standard waveform; the transient duration includes the time length of the high amplitude change of the voltage and the time length of the high amplitude change of the current; the spectrum amplitude includes the amplitude distribution of the voltage signal at different frequencies and the amplitude distribution of the current signal at different frequencies; the frequency ratio includes the amplitude ratio of the fundamental wave and the harmonic corresponding to the voltage signal and the amplitude ratio of the fundamental wave and the harmonic corresponding to the current signal; The method for extracting a fault feature vector from a fault data set comprises: The voltage standard waveform and the current standard waveform in the fault data set are respectively subjected to wavelet transformation to obtain the transient characteristics of the fault; the voltage standard waveform and the current standard waveform in the fault data set are respectively input into the trained feature extraction model to extract the fault feature data, which includes the fault waveform feature and the fault spectrum feature; wherein the feature extraction model includes A single feature extraction model, is the number of data in the fault feature data, and the single feature extraction model corresponds one-to-one to the data in the fault feature data; Each single feature extraction model is a deep neural network model, and The training process of each single feature extraction model is the same.

[0010] Further, the protection detection data is a circuit breaker trip signal; Methods for locating the fault location include: Fourier transform is performed on the voltage standard waveform and current standard waveform in the fault data set respectively, and each section of the voltage standard waveform and the current standard waveform is converted from the time domain to the frequency domain respectively, and the corresponding fundamental components are extracted respectively; the fundamental components corresponding to each section of the voltage standard waveform are marked as voltage fundamental components, and the fundamental components corresponding to each section of the current standard waveform are marked as current fundamental components; each voltage fundamental component is divided by the corresponding current fundamental component to obtain the corresponding impedance; the line of the transmission network corresponding to the protection detection data is marked as a fault line, and the impedance corresponding to the fault line is obtained and marked as the fault impedance; the number of fault impedances is counted and marked as the number of impedances; according to the number of impedances, the fault distance is calculated using the calculated impedance and the distance calculation model respectively, and the fault distance calculated using the calculated impedance is marked as the first distance, and the fault distance calculated using the distance calculation model is marked as the second distance; wherein the calculated impedance is the fault impedance used to calculate the fault distance, and the fault distance is the distance between the fault location and the fault recording device corresponding to the calculated impedance. The training process of the distance calculation model is consistent with the training process of the single feature extraction model, and both are deep neural network models; Get historical data, including data from different historical moments The distance error data includes calculation error data and model error data. is an integer greater than 1; wherein the calculated error data is the difference between the fault distance calculated by the calculated impedance and the corresponding actual distance, the model error data is the difference between the fault distance calculated by the distance calculation model and the corresponding actual distance, and the actual distance is the distance between the actual fault location and the fault recording device corresponding to the calculated impedance; the randomness corresponding to the calculated error data and the model error data is calculated respectively, and the weight coefficients corresponding to the calculated error data and the model error data are calculated respectively according to the randomness; the first distance is multiplied by the weight coefficient corresponding to the calculated error data, and the second distance is multiplied by the weight coefficient corresponding to the model error data to obtain the weighted distance, and the fault location is located according to the weighted distance.

[0011] Furthermore, the method of calculating the fault distance by respectively using the calculated impedance and the distance calculation model includes: If the number of impedances is equal to 2, a fault impedance is randomly used as the calculation impedance; the grid topology data is obtained, and according to the grid topology data, the total line length and unit impedance corresponding to the fault line are obtained, and the total line length is multiplied by the unit impedance to obtain the total line impedance; the calculated impedance is divided by the total line impedance, and then multiplied by the total line length to obtain the first distance; If the number of impedances is greater than 2, the fault recording devices at both ends of the fault line are marked as edge devices, the fault impedance corresponding to the edge device is used as the edge impedance, one edge impedance is randomly used as the calculation impedance, and the corresponding first distance is calculated; the fault recording device corresponding to the calculation impedance is marked as the calculation device, and the fault recording device not marked as the edge device is marked as the intermediate device; according to the power grid topology, the distance between each intermediate device and the calculation device is obtained and marked as the device distance; if there is a device distance less than or equal to twice the first distance, the corresponding intermediate device is marked as a recalculation device, the calculation impedance is replaced with the fault impedance corresponding to the recalculation device, and the corresponding first distance is recalculated; if all device distances are greater than twice the first distance, the first distance is not recalculated; If the number of impedances is equal to 2, the fault waveform characteristics, fault transient characteristics and fault spectrum characteristics corresponding to the calculation device are obtained from the fault feature vector and marked as the first positioning data; if the number of impedances is greater than 2, the fault waveform characteristics, fault transient characteristics and fault spectrum characteristics corresponding to the recalculation device are obtained from the fault feature vector and marked as the second positioning data; the first positioning data or the second positioning data is input into the trained distance calculation model to calculate the corresponding second distance.

[0012] Further, the method of calculating randomness includes: The number corresponding to each different numerical value in the error data is calculated and marked as the numerical number; the number corresponding to all numerical values ​​in the error data is calculated and marked as the total numerical number; each numerical number is divided by the total numerical number to obtain the numerical value probability; 2 is used as the base, and each numerical value probability is used as the true number in turn, and the information amount corresponding to each numerical value probability is calculated respectively; each information amount is multiplied by the corresponding numerical value probability to obtain the numerical contribution; each numerical contribution is added in turn, and then the inverse is taken to obtain the randomness corresponding to the error data; the calculation method of the randomness corresponding to the model error data is consistent with the calculation method of the randomness corresponding to the error data; Methods for calculating weight coefficients include: The randomness corresponding to the calculated error data is marked as the first randomness, and the randomness corresponding to the model error data is marked as the second randomness; the first randomness is added to the second randomness to obtain the total randomness; the first randomness is divided by the total randomness to obtain the weight coefficient corresponding to the calculated error data; the second randomness is divided by the total randomness to obtain the weight coefficient corresponding to the model error data.

[0013] Furthermore, the method for obtaining the fault geographical location includes: Import the power grid topology data into the geographic information system, annotate the coordinates of each fault recording device in the transmission network and mark them as device coordinates; obtain the device coordinates of the fault recording device corresponding to the calculated impedance and mark them as standard coordinates; generate a positioning circle with the standard coordinates as the center and the fault distance as the radius; mark the intersection of the positioning circle and the fault line as the fault point, count the number of fault points and mark them as the number of faults; if the number of faults is equal to 1, obtain the coordinates corresponding to the fault point from the geographic information system and use them as the fault geographical location; if the number of faults is greater than 1, obtain the device coordinates corresponding to the calculation device and mark them as the calculation coordinates; generate an auxiliary circle with the calculation coordinates as the center and the fault distance as the radius; mark the intersection of the auxiliary circle and the fault line as an auxiliary point; obtain the coordinates of each fault point and mark them as the first coordinate; obtain the coordinates of each auxiliary point and mark them as the second coordinate; calculate the Euclidean distance between each first coordinate and each second coordinate in turn and mark them as point distance; use the first coordinate corresponding to the smallest point distance as the fault geographical location.

[0014] Furthermore, the method for generating a fault diagnosis report includes: The fault recording device corresponding to each set of fault electrical quantity data is marked as a fault detection device; according to the power grid topology data, the line set between each fault detection device and the fault geographical location and the total line length corresponding to each line in each line set are obtained; the total line length corresponding to each line in each line set is added in sequence to obtain the influence range corresponding to each line set; each influence range is sorted from large to small to generate an influence sorting table; the influence range ranked first in the influence sorting table is used as the fault influence range; The fault feature vector and the fault impact range are used as impact data, and the impact data is input into the trained severity prediction model to predict the corresponding fault severity. The training process of the severity prediction model is consistent with the training process of the single feature extraction model, and both are deep neural network models. Generate a fault diagnosis report based on the fault impact scope, fault severity and fault geographical location.

[0015] A hybrid diagnosis system for power transmission network faults based on intelligent AI, used to implement the hybrid diagnosis method for power transmission network faults based on intelligent AI, comprising: Data acquisition module, used to collect Group fault electrical quantity data, is an integer greater than 1; A data processing module is used to perform standardization processing on each set of fault electrical quantity data and construct a fault data set; The feature extraction module is used to extract the fault feature vector from the fault data set by using wavelet transform and deep learning technology. The fault feature vector includes fault waveform feature, fault transient feature and fault spectrum feature. Fault location module, used to obtain protection detection data, fuse fault feature vectors and protection detection data, and use hybrid diagnosis strategy to quickly locate the fault location; The location correction module is used to obtain the power grid topology data, and combine the power grid topology data with the geographic information system to perform geographic coordinate correction on the fault location to obtain the fault geographical location; The risk assessment module is used to calculate the fault impact scope and fault severity according to the fault geographical location and fault feature vector, and generate a fault diagnosis report.

[0016] The technical effects and advantages of the hybrid diagnosis method and system for power transmission network faults based on intelligent AI of the present invention are as follows: By time-aligning the collected multiple sets of fault electrical quantity data, a unified fault data set is established, which effectively solves the problem of data time axis deviation caused by factors such as sampling rate and signal propagation delay; intelligent AI technology is used to realize the automatic extraction of various fault characteristic parameters, which can comprehensively characterize the characteristics of transmission network faults; protection detection data and fault feature vectors are integrated, and a hybrid diagnosis strategy is used to locate the fault position, improve the positioning accuracy in complex fault scenarios, and reduce the impact of errors; the fault position is corrected in combination with the power grid topology data and the geographic information system to determine the precise geographical location where the fault occurred; the fault impact range and severity are calculated based on the fault characteristics and geographical location, and the impact of the fault on the power grid is comprehensively evaluated, providing a more reliable fault diagnosis basis for power grid dispatchers; it can effectively improve the robustness and response speed of fault diagnosis, accurately locate the fault location, reduce the fault recovery time, and enhance the safety and stability of the transmission network, with significant economic and social benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a schematic diagram of a hybrid diagnosis system for power transmission network faults based on intelligent AI according to Embodiment 1 of the present invention; Figure 2 This is a flow chart of a hybrid diagnosis method for power transmission network faults based on intelligent AI according to Example 2 of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 creative work are within the scope of protection of the present invention.

[0019] Example 1

[0020] See also Figure 1 As shown, the present embodiment is a hybrid diagnosis system for power transmission network faults based on intelligent AI, including a data acquisition module, a data processing module, a feature extraction module, a fault location module, a position correction module and a risk assessment module; each module is connected by wire and / or wireless means to realize data transmission between modules.

[0021] Data acquisition module, used to collect Group fault electrical quantity data, is an integer greater than 1.

[0022] Fault electrical quantity data include voltage waveform data and current waveform data, which are obtained through the fault recording device installed on the transmission network; the voltage waveform data is the waveform of the voltage changing with time at each measuring point in the transmission network, which is used to analyze the voltage amplitude, phase angle, drop, etc.; the current waveform data is the waveform of the current changing with time at each measuring point in the transmission network, which is used to analyze the current amplitude, impact characteristics, short-circuit current, etc.

[0023] The data processing module is used to perform standardization processing on each set of fault electrical quantity data and construct a fault data set.

[0024] Methods for constructing fault datasets include: De-noising is performed on each group of fault electrical quantities to obtain preliminary processing data, which includes preliminary voltage processing data and preliminary current processing data. The purpose of de-noising is to remove noise caused by factors such as equipment error, environmental interference, and measurement error, and to improve data quality. De-noising methods include wavelet de-noising, median filtering, and Kalman filtering. Count the number of data in each group of preliminary voltage processing data respectively and mark them as data number; compare each data number respectively, mark the data number with the largest value as the maximum number, and mark the preliminary voltage processing data corresponding to the maximum number as voltage standard data; mark the preliminary current processing data with the same fault recording device as the voltage standard data as the current standard data; mark the preliminary voltage processing data not marked as voltage standard data as voltage data to be aligned, and mark the preliminary current processing data not marked as current standard data as current data to be aligned; Each set of voltage data to be aligned is time-aligned with the voltage standard data to obtain voltage alignment data; each set of current data to be aligned is time-aligned with the current standard data to obtain current alignment data; a fault data set is constructed based on each set of voltage alignment data, current alignment data, voltage standard data and current standard data; it should be understood that in the analysis of power transmission network faults, the data collected by different recording devices have deviations on the time axis due to factors such as sampling rate and signal propagation delay, and therefore need to be synchronized and the time axis of the waveform adjusted to ensure that the data at each measurement point can be correctly matched.

[0025] The method for time-aligning the voltage data to be aligned with the voltage standard data is consistent with the method for time-aligning the current data to be aligned with the current standard data; the method for time-aligning the voltage data to be aligned with the voltage standard data includes: According to the time sequence, the voltages in the voltage data to be aligned and the voltage standard data are obtained respectively, and the corresponding voltage sequences are constructed respectively; the Euclidean distance between every two voltages in the two voltage sequences is calculated respectively, and marked as voltage distance; a distance matrix is ​​constructed according to the voltage distance, and the elements in the distance matrix correspond to the voltage distance one by one; the calculation method of the Euclidean distance is a prior art and will not be described in detail here; according to the distance matrix, a cumulative distance matrix is ​​constructed, and the size of the cumulative distance matrix is ​​consistent with the size of the distance matrix; the elements in the first row and the first column of the cumulative distance matrix are marked as starting elements, the elements in the first row of the cumulative distance matrix are marked as head elements, the elements in the first column of the cumulative distance matrix are marked as index elements, and the elements in the cumulative distance matrix that are not marked as starting elements, head elements and index elements are all marked as data elements; The starting point element in the cumulative distance matrix is ​​equal to the starting point element in the distance matrix. Each head element in the cumulative distance matrix is ​​the previous head element plus the corresponding element in the distance matrix. Each index element in the cumulative distance matrix is ​​the previous index element plus the corresponding element in the distance matrix. Each data element in the cumulative distance matrix is ​​the minimum value of the corresponding adjacent elements plus the corresponding element in the distance matrix. The adjacent elements corresponding to the data element include the element located to the left of the data element, the element located above the data element, and the element located to the upper left of the data element in the cumulative distance matrix. Starting from the element corresponding to the lower right corner of the cumulative distance matrix, backtrack to the starting element of the cumulative distance matrix to obtain the optimal path; the backtracking process is: mark the element backtracked to as the current element, select the minimum value of the adjacent elements corresponding to the current element in the cumulative distance matrix as the backtracking element, and update the current element to the backtracking element; according to the optimal path, establish a voltage matching relationship, and the voltage matching relationship is the corresponding relationship between the voltage in the voltage data to be aligned and the voltage in the voltage standard data; according to the voltage matching relationship, interpolate the voltage data to be aligned so that the voltage data to be aligned is aligned with the voltage standard data on the time axis; the method for interpolating the voltage data to be aligned is: for a voltage in the data to be aligned in the optimal path, the voltage in the voltage standard data is matched. When a voltage is generated, the corresponding voltage is marked as the interpolation voltage, and a linear interpolation method is used at the interpolation voltage to generate A new voltage, is an integer greater than 1; the linear interpolation method is a prior art and will not be described in detail here.

[0026] Exemplarily, the voltage sequences corresponding to the voltage data to be aligned are 101 and 106, and the voltage sequences corresponding to the voltage standard data are 100, 102, and 105; therefore, the distance matrix is: , the cumulative distance matrix: ; The optimal path includes 3, 2, and 1 in the cumulative distance matrix, that is, the optimal path is ; The voltage matching relationship is that the first voltage in the voltage data to be aligned corresponds to the first voltage in the voltage standard data, the first voltage in the voltage data to be aligned corresponds to the second voltage in the voltage standard data, and the second voltage in the voltage data to be aligned corresponds to the third voltage in the voltage standard data; since the first voltage in the voltage data to be aligned corresponds to two voltages in the voltage standard data, the first voltage in the voltage data to be aligned is marked as an interpolated voltage, and a new voltage 103 is generated after the interpolated voltage using a linear interpolation method.

[0027] The feature extraction module is used to extract fault feature vectors from the fault data set by using wavelet transform and deep learning technology. The fault feature vectors include fault waveform features, fault transient features and fault spectrum features.

[0028] The fault feature vector helps to fully characterize the characteristics and properties of transmission network faults, thereby achieving accurate fault location and identification; Fault waveform characteristics are the main characteristics of the voltage waveform and current waveform when a transmission network fault occurs, which can reflect the basic form and change trend of the fault signal; fault waveform characteristics include amplitude characteristics and waveform change rate; fault transient characteristics are the sudden change or transient response of current and voltage at the moment of transmission network fault or within a short period of time after the occurrence; fault transient characteristics include transient amplitude and transient duration; fault spectrum characteristics are the frequency distribution and energy distribution of voltage and current during the transmission network fault process, and fault spectrum characteristics include spectrum amplitude and frequency ratio; The voltage waveform data after standardization are all marked as voltage standard waveforms, and the current waveform data are all marked as current standard waveforms; the amplitude characteristics include the maximum amplitude (i.e., the value of the highest point in the waveform) and the minimum amplitude (i.e., the value of the lowest point in the waveform) of the voltage standard waveform, and the maximum amplitude and minimum amplitude of the current standard waveform; the waveform change rate includes the maximum instantaneous slope of the voltage standard waveform (i.e., the maximum rate of change of voltage in unit time) and the maximum instantaneous slope of the current standard waveform; the transient amplitude includes the maximum instantaneous deviation value of the voltage standard waveform (i.e., the maximum deviation of the waveform from its average value) and the maximum instantaneous deviation value of the current standard waveform; the transient duration includes the time length of high amplitude change of voltage (i.e., the duration of the amplitude deviation of the waveform from the preset normal threshold, the normal threshold is pre-set by technical personnel in this field according to actual conditions) and the time length of high amplitude change of current; the spectrum amplitude includes the amplitude distribution of the voltage signal at different frequencies and the amplitude distribution of the current signal at different frequencies; the frequency ratio includes the amplitude ratio of the fundamental wave to the harmonic corresponding to the voltage signal and the amplitude ratio of the fundamental wave to the harmonic corresponding to the current signal.

[0029] Methods for extracting fault feature vectors from fault data sets include: The voltage standard waveform and the current standard waveform in the fault data set are respectively subjected to wavelet transformation to obtain the transient characteristics of the fault; the voltage standard waveform and the current standard waveform in the fault data set are respectively input into the trained feature extraction model to extract the fault feature data, which includes the fault waveform feature and the fault spectrum feature; wherein the feature extraction model includes A single feature extraction model, is the number of data in the fault feature data, and the single feature extraction model corresponds to the data in the fault feature data one by one, that is, a single feature extraction model extracts one type of data in the fault feature data. is 5, namely, including the maximum amplitude, the minimum amplitude, the maximum instantaneous slope, the amplitude distribution at different frequencies, and the amplitude ratio of the fundamental wave to the harmonics; illustratively, the single feature extraction model corresponds to the maximum amplitude in the fault feature data, so the voltage standard waveform is input into the single feature extraction model to extract the maximum amplitude of the voltage standard waveform, and the current standard waveform is input into the single feature extraction model to extract the maximum amplitude of the current standard waveform; Each single feature extraction model is a deep neural network model, and The training process of each single feature extraction model is the same.

[0030] The training process of the maximum amplitude corresponding single feature extraction model includes: Pre-collection There are different standard waveforms, the standard waveform is a voltage standard waveform or a current standard waveform. The corresponding maximum amplitude is set for each standard waveform. is an integer greater than 1, and the standard waveform and the corresponding maximum amplitude are converted into a corresponding set of feature vectors; the maximum amplitude corresponding to the standard waveform is collected by technicians in the process of extracting fault feature data historically. Different standard waveforms are analyzed in turn according to the actual situation, and the corresponding maximum amplitude is extracted from each standard waveform. Set the corresponding maximum amplitudes for different standard waveforms in turn; Each set of feature vectors is used as the input of the single feature extraction model. The single feature extraction model uses a set of predicted maximum amplitudes corresponding to each set of standard waveforms as output, and the actual maximum amplitude corresponding to each set of standard waveforms as the prediction target. The actual maximum amplitude is the preset maximum amplitude corresponding to the standard waveform. The training goal is to minimize the sum of the prediction errors of all standard waveforms. The calculation formula of the prediction error is: ,in is the prediction error, is the group number of the characteristic vector corresponding to the standard waveform, For the The predicted maximum amplitude corresponding to the standard waveform of the group, For the The actual maximum amplitude corresponding to the standard waveform of the group; the single feature extraction model is trained until the sum of the prediction errors reaches convergence and the training is stopped.

[0031] The fault location module is used to obtain protection detection data, fuse fault feature vectors and protection detection data, and use a hybrid diagnosis strategy to quickly locate the fault location.

[0032] The protection detection data is the circuit breaker tripping signal, which is obtained through the electrical protection device (circuit breaker) installed on the transmission network; the circuit breaker tripping signal is the signal sent by the circuit breaker when performing the disconnection operation, which is used to indicate that the fault line in the transmission network has been cut off, that is, effectively isolating the fault in the transmission network, preventing the fault from spreading to other parts, and ensuring the normal operation of other lines.

[0033] Methods for locating the fault location include: Fourier transform is performed on the standard voltage waveform and the standard current waveform in the fault data set respectively, each section of the standard voltage waveform and the standard current waveform is converted from the time domain to the frequency domain respectively, and the corresponding fundamental components are extracted respectively; the fundamental components corresponding to each section of the standard voltage waveform are marked as voltage fundamental components, and the fundamental components corresponding to each section of the standard current waveform are marked as current fundamental components; each voltage fundamental component is divided by the corresponding current fundamental component to obtain the corresponding impedance; the line of the transmission network corresponding to the protection detection data (i.e., the transmission line) is marked as a fault line, the impedance corresponding to the fault line is obtained, and it is marked as the fault impedance; the fault statistics are performed. The number of barrier impedances is calculated and marked as the impedance number; Fourier transform is a prior art, and the specific process is not described in detail here; according to the impedance number, the fault distance is calculated by using the calculated impedance and the distance calculation model respectively, the fault distance calculated by using the calculated impedance is marked as the first distance, and the fault distance calculated by using the distance calculation model is marked as the second distance; wherein the calculated impedance is the fault impedance used to calculate the fault distance, the fault distance is the distance between the fault location and the fault recording device corresponding to the calculated impedance, the training process of the distance calculation model is consistent with the training process of the single feature extraction model, and both are deep neural network models; Get historical data, including data from different historical moments The distance error data includes calculation error data and model error data. is an integer greater than 1; wherein the calculated error data is the difference between the fault distance calculated by the calculated impedance and the corresponding actual distance, the model error data is the difference between the fault distance calculated by the distance calculation model and the corresponding actual distance, and the actual distance is the distance between the actual fault location and the fault recording device corresponding to the calculated impedance; the randomness corresponding to the calculated error data and the model error data is calculated respectively, and the weight coefficients corresponding to the calculated error data and the model error data are calculated respectively according to the randomness; the first distance is multiplied by the weight coefficient corresponding to the calculated error data, and the second distance is multiplied by the weight coefficient corresponding to the model error data to obtain the weighted distance, and the fault location is located according to the weighted distance.

[0034] Methods for calculating the fault distance using the impedance calculation model and the distance calculation model include: If the number of impedances is equal to 2, a fault impedance is randomly used as the calculated impedance; the grid topology data is obtained, which is information describing the grid structure, line connection relationship and equipment parameters, and reflects the physical and electrical connection relationship between substations, transmission lines, loads, switchgear, generators and other components in the transmission network; the grid topology data is obtained through the SCADA system of the power grid company, local power supply company or power dispatching center; according to the grid topology data, the total line length and unit impedance corresponding to the fault line are obtained, and the total line length is multiplied by the unit impedance to obtain the total line impedance; the calculated impedance is divided by the total line impedance, and then multiplied by the total line length to obtain the first distance; If the number of impedances is greater than 2, the fault recording devices located at both ends of the fault line are marked as edge devices, the fault impedance corresponding to the edge device is used as the edge impedance, one edge impedance is randomly used as the calculation impedance, and the corresponding first distance is calculated; the fault recording device corresponding to the calculation impedance is marked as the calculation device, and the fault recording device not marked as the edge device is marked as the intermediate device; according to the power grid topology, the distance between each intermediate device and the calculation device is obtained and marked as the device distance; if there is a device distance less than or equal to twice the first distance, the corresponding intermediate device is marked as a recalculation device, the calculation impedance is replaced with the fault impedance corresponding to the recalculation device, and the corresponding first distance is recalculated; if all device distances are greater than twice the first distance, the first distance is not recalculated.

[0035] It should be noted that since at least two fault recording devices are configured on a transmission line, there is no situation where the impedance number is less than 2; the reason for recalculating the fault distance is that when there is a device distance less than or equal to twice the fault distance, it means that the distance between the corresponding intermediate device and the fault location is closer, and the corresponding impedance can better reflect the electrical characteristics of the fault location. Therefore, the recalculated fault distance can more accurately locate the actual fault location.

[0036] If the number of impedances is equal to 2, the fault waveform characteristics, fault transient characteristics and fault spectrum characteristics corresponding to the calculation device are obtained from the fault feature vector and marked as the first positioning data; if the number of impedances is greater than 2, the fault waveform characteristics, fault transient characteristics and fault spectrum characteristics corresponding to the recalculation device are obtained from the fault feature vector and marked as the second positioning data; the first positioning data or the second positioning data is input into the trained distance calculation model to calculate the corresponding second distance.

[0037] Methods for calculating randomness include: The number corresponding to each different numerical value in the calculated error data is counted and marked as the numerical number; the number corresponding to all numerical values ​​in the calculated error data is counted and marked as the total numerical number; each numerical number is divided by the total numerical number to obtain the numerical value probability; 2 is used as the base and each numerical value probability is used as the true number in turn to calculate the amount of information corresponding to each numerical value probability; each amount of information is multiplied by the corresponding numerical value probability to obtain the numerical contribution; each numerical contribution is added in turn and then the inverse is taken to obtain the randomness corresponding to the calculated error data; the calculation method for the randomness corresponding to the model error data is consistent with the calculation method for the randomness corresponding to the calculated error data.

[0038] Methods for calculating weight coefficients include: The randomness corresponding to the calculated error data is marked as the first randomness, and the randomness corresponding to the model error data is marked as the second randomness; the first randomness is added to the second randomness to obtain the total randomness; the first randomness is divided by the total randomness to obtain the weight coefficient corresponding to the calculated error data; the second randomness is divided by the total randomness to obtain the weight coefficient corresponding to the model error data.

[0039] It should be understood that when a fault occurs in the transmission network, electrical disturbances will propagate throughout the transmission network, causing the voltage and current of the lines adjacent to the faulty line to fluctuate, thereby triggering the fault recording devices on these lines; however, electrical protection devices can usually correctly identify that only the faulty line needs to be disconnected, that is, only the electrical protection device on the faulty line will generate a circuit breaker tripping signal; therefore, the fault line in the transmission network cannot be accurately determined by relying solely on the fault characteristic vector, and protection detection data must be combined to accurately determine the faulty line.

[0040] The location correction module is used to obtain the power grid topology data, combine the power grid topology data with the geographic information system, perform geographic coordinate correction on the fault location, and obtain the fault geographical location.

[0041] Methods for obtaining the fault geographic location include: Import the grid topology data into the geographic information system, annotate the coordinates of each fault recording device in the transmission network, and mark them as device coordinates; the geographic information system is a system used to store, manage, analyze, process and visualize geographic data, which can combine spatial data (such as maps, geographic coordinates, terrain, etc.) with non-spatial data (such as transmission lines, fault recording devices, etc.), which helps to obtain relevant information about the geographic location; obtain the device coordinates of the fault recording device corresponding to the calculated impedance, and mark them as standard coordinates; generate a positioning circle with the standard coordinates as the center and the fault distance as the radius; mark the intersection of the positioning circle and the fault line as the fault point, count the number of fault points, and mark them as the number of faults; if the number of faults is equal to 1, The coordinates corresponding to the fault point are obtained from the system and used as the fault geographical location; if the number of faults is greater than 1, the device coordinates corresponding to the computing device are obtained and marked as the calculated coordinates; an auxiliary circle is generated with the calculated coordinates as the center and the fault distance as the radius; the intersection of the auxiliary circle and the fault line is marked as an auxiliary point; the coordinates of each fault point are obtained and marked as the first coordinates; the coordinates of each auxiliary point are obtained and marked as the second coordinates; the Euclidean distance between each first coordinate and each second coordinate is calculated in turn and marked as the point distance. The calculation method of the Euclidean distance is a prior art and the specific process is not described in detail here; the first coordinate corresponding to the minimum point distance is used as the fault geographical location; since there is a fault location, there is no situation where the number of faults is less than 1.

[0042] The risk assessment module is used to calculate the fault impact scope and fault severity according to the fault geographical location and fault feature vector, and generate a fault diagnosis report.

[0043] Methods for generating a troubleshooting report include: The fault recording device corresponding to each set of fault electrical quantity data is marked as a fault detection device; according to the power grid topology data, the line set between each fault detection device and the fault geographical location and the total line length corresponding to each line in each line set are obtained; the total line length corresponding to each line in each line set is added in sequence to obtain the influence range corresponding to each line set; each influence range is sorted from large to small to generate an influence sorting table; the influence range ranked first in the influence sorting table is used as the fault influence range; The fault feature vector and the fault impact range are used as impact data, and the impact data is input into the trained severity prediction model to predict the corresponding fault severity. The training process of the severity prediction model is consistent with the training process of the single feature extraction model, and both are deep neural network models. Generate a fault diagnosis report based on the fault impact scope, fault severity and fault geographical location.

[0044] This embodiment establishes a unified fault data set by time-aligning multiple groups of collected fault electrical quantity data, effectively solving the problem of data time axis deviation caused by factors such as sampling rate and signal propagation delay; adopts intelligent AI technology to realize automatic extraction of multiple fault characteristic parameters, and can comprehensively characterize the characteristics of transmission network faults; integrates protection detection data and fault feature vectors, adopts a hybrid diagnosis strategy to locate the fault position, improves the positioning accuracy in complex fault scenarios, and reduces the impact of errors; combines power grid topology data and geographic information systems to correct the fault position and determine the precise geographical location where the fault occurred; calculates the fault impact range and severity based on fault characteristics and geographical location, comprehensively evaluates the impact of the fault on the power grid, and provides a more reliable fault diagnosis basis for power grid dispatchers; can effectively improve the robustness and response speed of fault diagnosis, accurately locate the fault location, reduce fault recovery time, enhance the safety and stability of the transmission network, and has significant economic and social benefits.

[0045] Example 2

[0046] See also Figure 2 As shown, the part not described in detail in this embodiment is described in Example 1, and a hybrid diagnosis method for power transmission network faults based on intelligent AI is provided, the method comprising: collection Group fault electrical quantity data, is an integer greater than 1; Each set of fault electrical quantity data is standardized to construct a fault data set; By using wavelet transform and deep learning technology, the fault feature vector is extracted from the fault data set. The fault feature vector includes fault waveform features, fault transient features and fault spectrum features. Obtain protection detection data, fuse fault feature vectors and protection detection data, and use a hybrid diagnosis strategy to quickly locate the fault location; Obtain the grid topology data, combine the grid topology data with the geographic information system, perform geographic coordinate correction on the fault location, and obtain the fault geographical location; According to the fault geographical location and fault feature vector, the fault impact range and fault severity are calculated, and a fault diagnosis report is generated.

[0047] Example 3

[0048] The present application also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memory stores a computer-readable code, and when the computer-readable code is executed by one or more processors, it can execute the above-mentioned hybrid diagnosis method for power transmission network faults based on intelligent AI.

[0049] The method or system according to the implementation mode of the present application can also be implemented with the help of the architecture of the electronic device shown in the present application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to the network, input / output, a hard disk, etc. A storage device in the electronic device, such as a ROM or a hard disk, can store a hybrid diagnosis method for power transmission network faults based on intelligent AI provided in the present application. Furthermore, the electronic device may also include a user interface. Of course, the architecture shown in the present application is only exemplary. When implementing different devices, one or more components in the electronic device shown in the present application may be omitted according to actual needs.

[0050] Example 4

[0051] One embodiment of the present application discloses a computer-readable storage medium. Computer-readable instructions are stored on the computer-readable storage medium. When the computer-readable instructions are executed by the processor, a hybrid diagnosis method for power transmission network faults based on intelligent AI according to an embodiment of the present application described with reference to the above figures can be executed. The storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0052] In addition, according to the implementation of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided in the present application, for example: a hybrid diagnosis method for power transmission network faults based on intelligent AI. When the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0053] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0054] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A hybrid diagnosis method for power transmission network faults based on intelligent AI, characterized in that: include: collection Group fault electrical quantity data, is an integer greater than 1; Each set of fault electrical quantity data is standardized to construct a fault data set; Using wavelet transform and deep learning technology, fault feature vectors are extracted from fault data sets. The fault feature vector includes fault waveform feature, fault transient feature and fault spectrum feature; Obtain protection detection data, fuse fault feature vectors and protection detection data, and use a hybrid diagnosis strategy to quickly locate the fault location; Obtain the grid topology data, combine the grid topology data with the geographic information system, perform geographic coordinate correction on the fault location, and obtain the fault geographical location; According to the fault geographical location and fault feature vector, the fault impact range and fault severity are calculated, and a fault diagnosis report is generated.

2. The hybrid diagnosis method for power transmission network faults based on intelligent AI according to claim 1 is characterized in that: The fault electrical quantity data includes voltage waveform data and current waveform data; The method for constructing a fault data set comprises: De-noising is performed on each group of fault electrical quantities to obtain preliminary processing data, which includes preliminary voltage processing data and preliminary current processing data; Count the number of data in each group of preliminary voltage processing data respectively and mark them as data number; compare each data number respectively, mark the data number with the largest value as the maximum number, and mark the preliminary voltage processing data corresponding to the maximum number as voltage standard data; mark the preliminary current processing data with the same fault recording device as the voltage standard data as the current standard data; mark the preliminary voltage processing data not marked as voltage standard data as voltage data to be aligned, and mark the preliminary current processing data not marked as current standard data as current data to be aligned; Each group of voltage data to be aligned is time-aligned with the voltage standard data to obtain voltage alignment data; each group of current data to be aligned is time-aligned with the current standard data to obtain current alignment data; a fault data set is constructed according to each group of voltage alignment data, current alignment data, voltage standard data and current standard data; wherein, the method of time-aligning the voltage data to be aligned with the voltage standard data is consistent with the method of time-aligning the current data to be aligned with the current standard data.

3. The hybrid diagnosis method for power transmission network faults based on intelligent AI according to claim 2 is characterized in that: The method for time-aligning the voltage data to be aligned with the voltage standard data includes: According to the time sequence, the voltages in the voltage data to be aligned and the voltage standard data are obtained respectively, and the corresponding voltage sequences are constructed respectively; the Euclidean distance between every two voltages in the two voltage sequences is calculated respectively, and marked as voltage distance; a distance matrix is ​​constructed according to the voltage distance, and the elements in the distance matrix correspond to the voltage distance one by one; according to the distance matrix, a cumulative distance matrix is ​​constructed, and the size of the cumulative distance matrix is ​​consistent with the size of the distance matrix; the elements in the first row and the first column of the cumulative distance matrix are marked as starting elements, the elements in the first row of the cumulative distance matrix are marked as head elements, the elements in the first column of the cumulative distance matrix are marked as index elements, and the elements in the cumulative distance matrix that are not marked as starting elements, head elements and index elements are all marked as data elements; The starting point element in the cumulative distance matrix is ​​equal to the starting point element in the distance matrix. Each head element in the cumulative distance matrix is ​​the previous head element plus the corresponding element in the distance matrix. Each index element in the cumulative distance matrix is ​​the previous index element plus the corresponding element in the distance matrix. Each data element in the cumulative distance matrix is ​​the minimum value of the corresponding adjacent elements plus the corresponding element in the distance matrix. The adjacent elements corresponding to the data element include the element located to the left of the data element, the element located above the data element, and the element located to the upper left of the data element in the cumulative distance matrix. Starting from the element corresponding to the lower right corner of the cumulative distance matrix, backtrack to the starting element of the cumulative distance matrix to obtain the optimal path; the backtracking process is: mark the element backtracked to as the current element, select the minimum value among the adjacent elements corresponding to the current element in the cumulative distance matrix as the backtracking element, and update the current element as the backtracking element; according to the optimal path, establish a voltage matching relationship, which is the corresponding relationship between the voltage in the voltage data to be aligned and the voltage in the voltage standard data; according to the voltage matching relationship, interpolate the voltage data to be aligned; the method for interpolating the voltage data to be aligned is: for a voltage in the data to be aligned in the optimal path, the voltage in the voltage standard data is corresponding to the voltage in the voltage standard data. When a voltage is generated, the corresponding voltage is marked as the interpolation voltage, and a linear interpolation method is used at the interpolation voltage to generate A new voltage, is an integer greater than 1.

4. The hybrid diagnosis method for power transmission network faults based on intelligent AI according to claim 3 is characterized in that: The fault waveform characteristics include amplitude characteristics and waveform change rate; the fault transient characteristics include transient amplitude and transient duration; The fault spectrum characteristics include spectrum amplitude and frequency ratio; The voltage waveform data after the standardization process are all marked as voltage standard waveforms, and the current waveform data are all marked as current standard waveforms; The amplitude characteristics include the maximum amplitude and minimum amplitude of the voltage standard waveform, and the maximum amplitude and minimum amplitude of the current standard waveform; the waveform change rate includes the maximum instantaneous slope of the voltage standard waveform and the maximum instantaneous slope of the current standard waveform; the transient amplitude includes the maximum instantaneous deviation value of the voltage standard waveform and the maximum instantaneous deviation value of the current standard waveform; the transient duration includes the duration of the high amplitude change of the voltage and the duration of the high amplitude change of the current; The spectrum amplitude includes the amplitude distribution of the voltage signal at different frequencies and the amplitude distribution of the current signal at different frequencies; the frequency ratio includes the amplitude ratio of the fundamental wave and the harmonic corresponding to the voltage signal and the amplitude ratio of the fundamental wave and the harmonic corresponding to the current signal; The method for extracting a fault feature vector from a fault data set comprises: The voltage standard waveform and the current standard waveform in the fault data set are respectively subjected to wavelet transformation to obtain the transient characteristics of the fault; the voltage standard waveform and the current standard waveform in the fault data set are respectively input into the trained feature extraction model to extract the fault feature data, which includes the fault waveform feature and the fault spectrum feature; wherein the feature extraction model includes A single feature extraction model, is the number of data in the fault feature data, and the single feature extraction model corresponds one-to-one to the data in the fault feature data; Each single feature extraction model is a deep neural network model, and The training process of each single feature extraction model is the same.

5. The hybrid diagnosis method for power transmission network faults based on intelligent AI according to claim 4 is characterized in that: The protection detection data is a circuit breaker trip signal; Methods for locating the fault location include: Fourier transform is performed on the voltage standard waveform and current standard waveform in the fault data set respectively, and each section of the voltage standard waveform and the current standard waveform is converted from the time domain to the frequency domain respectively, and the corresponding fundamental components are extracted respectively; the fundamental components corresponding to each section of the voltage standard waveform are marked as voltage fundamental components, and the fundamental components corresponding to each section of the current standard waveform are marked as current fundamental components; each voltage fundamental component is divided by the corresponding current fundamental component to obtain the corresponding impedance; the line of the transmission network corresponding to the protection detection data is marked as a fault line, and the impedance corresponding to the fault line is obtained and marked as the fault impedance; the number of fault impedances is counted and marked as the number of impedances; according to the number of impedances, the fault distance is calculated using the calculated impedance and the distance calculation model respectively, and the fault distance calculated using the calculated impedance is marked as the first distance, and the fault distance calculated using the distance calculation model is marked as the second distance; wherein the calculated impedance is the fault impedance used to calculate the fault distance, and the fault distance is the distance between the fault location and the fault recording device corresponding to the calculated impedance. The training process of the distance calculation model is consistent with the training process of the single feature extraction model, and both are deep neural network models; Get historical data, including data from different historical moments The distance error data includes calculation error data and model error data. is an integer greater than 1; wherein the calculated error data is the difference between the fault distance calculated by the calculated impedance and the corresponding actual distance, the model error data is the difference between the fault distance calculated by the distance calculation model and the corresponding actual distance, and the actual distance is the distance between the actual fault location and the fault recording device corresponding to the calculated impedance; the randomness corresponding to the calculated error data and the model error data is calculated respectively, and the weight coefficients corresponding to the calculated error data and the model error data are calculated respectively according to the randomness; the first distance is multiplied by the weight coefficient corresponding to the calculated error data, and the second distance is multiplied by the weight coefficient corresponding to the model error data to obtain the weighted distance, and the fault location is located according to the weighted distance.

6. The hybrid diagnosis method for power transmission network faults based on intelligent AI according to claim 5 is characterized in that: The method for calculating the fault distance by respectively using the calculated impedance and the distance calculation model comprises: If the number of impedances is equal to 2, a fault impedance is randomly used as the calculation impedance; the grid topology data is obtained, and according to the grid topology data, the total line length and unit impedance corresponding to the fault line are obtained, and the total line length is multiplied by the unit impedance to obtain the total line impedance; the calculated impedance is divided by the total line impedance, and then multiplied by the total line length to obtain the first distance; If the number of impedances is greater than 2, the fault recording devices at both ends of the fault line are marked as edge devices, the fault impedance corresponding to the edge device is used as the edge impedance, one edge impedance is randomly used as the calculation impedance, and the corresponding first distance is calculated; the fault recording device corresponding to the calculation impedance is marked as the calculation device, and the fault recording device not marked as the edge device is marked as the intermediate device; according to the power grid topology, the distance between each intermediate device and the calculation device is obtained and marked as the device distance; if there is a device distance less than or equal to twice the first distance, the corresponding intermediate device is marked as a recalculation device, the calculation impedance is replaced with the fault impedance corresponding to the recalculation device, and the corresponding first distance is recalculated; if all device distances are greater than twice the first distance, the first distance is not recalculated; If the number of impedances is equal to 2, the fault waveform characteristics, fault transient characteristics and fault spectrum characteristics corresponding to the calculation device are obtained from the fault feature vector and marked as the first positioning data; if the number of impedances is greater than 2, the fault waveform characteristics, fault transient characteristics and fault spectrum characteristics corresponding to the recalculation device are obtained from the fault feature vector and marked as the second positioning data; the first positioning data or the second positioning data is input into the trained distance calculation model to calculate the corresponding second distance.

7. The hybrid diagnosis method for power transmission network faults based on intelligent AI according to claim 6 is characterized in that: Methods for calculating randomness include: The number corresponding to each different numerical value in the error data is calculated and marked as the numerical number; the number corresponding to all numerical values ​​in the error data is calculated and marked as the total numerical number; each numerical number is divided by the total numerical number to obtain the numerical value probability; 2 is used as the base, and each numerical value probability is used as the true number in turn, and the information amount corresponding to each numerical value probability is calculated respectively; each information amount is multiplied by the corresponding numerical value probability to obtain the numerical contribution; each numerical contribution is added in turn, and then the inverse is taken to obtain the randomness corresponding to the error data; the calculation method of the randomness corresponding to the model error data is consistent with the calculation method of the randomness corresponding to the error data; Methods for calculating weight coefficients include: The randomness corresponding to the calculated error data is marked as the first randomness, and the randomness corresponding to the model error data is marked as the second randomness; the first randomness is added to the second randomness to obtain the total randomness; the first randomness is divided by the total randomness to obtain the weight coefficient corresponding to the calculated error data; the second randomness is divided by the total randomness to obtain the weight coefficient corresponding to the model error data.

8. The hybrid diagnosis method for power transmission network faults based on intelligent AI according to claim 7 is characterized in that: The method for obtaining the fault geographical location includes: Import the power grid topology data into the geographic information system, annotate the coordinates of each fault recording device in the transmission network and mark them as device coordinates; obtain the device coordinates of the fault recording device corresponding to the calculated impedance and mark them as standard coordinates; generate a positioning circle with the standard coordinates as the center and the fault distance as the radius; mark the intersection of the positioning circle and the fault line as the fault point, count the number of fault points and mark them as the number of faults; if the number of faults is equal to 1, obtain the coordinates corresponding to the fault point from the geographic information system and use them as the fault geographical location; if the number of faults is greater than 1, obtain the device coordinates corresponding to the calculation device and mark them as the calculation coordinates; generate an auxiliary circle with the calculation coordinates as the center and the fault distance as the radius; mark the intersection of the auxiliary circle and the fault line as an auxiliary point; obtain the coordinates of each fault point and mark them as the first coordinate; obtain the coordinates of each auxiliary point and mark them as the second coordinate; calculate the Euclidean distance between each first coordinate and each second coordinate in turn and mark them as point distance; use the first coordinate corresponding to the smallest point distance as the fault geographical location.

9. The hybrid diagnosis method for power transmission network faults based on intelligent AI according to claim 8 is characterized in that: Methods for generating a troubleshooting report include: The fault recording device corresponding to each set of fault electrical quantity data is marked as a fault detection device; according to the power grid topology data, the line set between each fault detection device and the fault geographical location and the total line length corresponding to each line in each line set are obtained; the total line length corresponding to each line in each line set is added in sequence to obtain the influence range corresponding to each line set; each influence range is sorted from large to small to generate an influence sorting table; the influence range ranked first in the influence sorting table is used as the fault influence range; The fault feature vector and the fault impact range are used as impact data, and the impact data is input into the trained severity prediction model to predict the corresponding fault severity. The training process of the severity prediction model is consistent with the training process of the single feature extraction model, and both are deep neural network models. Generate a fault diagnosis report based on the fault impact scope, fault severity and fault geographical location.

10. A hybrid diagnosis system for power transmission network faults based on intelligent AI, used to implement the hybrid diagnosis method for power transmission network faults based on intelligent AI as claimed in any one of claims 1 to 9, characterized in that: include: Data acquisition module, used to collect Group fault electrical quantity data, is an integer greater than 1; A data processing module is used to perform standardization processing on each set of fault electrical quantity data and construct a fault data set; The feature extraction module is used to extract fault feature vectors from fault data sets using wavelet transform and deep learning technology. The fault feature vector includes fault waveform feature, fault transient feature and fault spectrum feature; Fault location module, used to obtain protection detection data, fuse fault feature vectors and protection detection data, and use hybrid diagnosis strategy to quickly locate the fault location; The location correction module is used to obtain the power grid topology data, and combine the power grid topology data with the geographic information system to perform geographic coordinate correction on the fault location to obtain the fault geographical location; The risk assessment module is used to calculate the fault impact scope and fault severity according to the fault geographical location and fault feature vector, and generate a fault diagnosis report.

Images