Method and system for identifying transient disturbance fragment of fault recording data

Abstract

The invention discloses a method and a system for identifying a transient disturbance fragment of fault recording data. The method comprises the following steps: collecting the fault recording data; constructing a recording data set, and performing integrity detection on the data set; screening and repairing the recording data set, and carrying out transient disturbance identification on fault recording data; and constructing a random forest model, and calibrating disturbance data fragments in the fault recording data. According to the method for identifying the transient disturbance fragment of the fault recording data provided by the invention, adaptive channel selection and file completion are realized by using data set prior information, filtering structure parameters are adaptively selected based on the sampling rate, disturbance waveform characteristics in the recording data set are analyzed, the disturbance data fragment can be positioned and identified, and the fault recording data can be identified. And data support is provided for a load side.

Description

Technical Field

[0001] This invention relates to the field of power system technology, specifically to a method and system for identifying transient disturbance segments in fault recording data. Background Technology

[0002] Compared to traditional power systems, new power systems present significant challenges to operation and regulation due to the high proportion of renewable energy integration and the electrification of the load side. Digital simulation is a crucial tool for studying power systems, and its effectiveness is directly influenced by the effectiveness of the load model.

[0003] Utilizing a large sample of measured data is a crucial foundational task for overall load model identification and subsequent simulation reproduction. The widespread application of fault recording devices on the user side provides an important source of measured data. Fault recorders offer advantages in load modeling, including no need for external devices, cost-effectiveness, and high sampling rate and accuracy. However, fault record files contain a significant amount of data unusable for load model identification, such as false starts, non-fault disturbance data, and switching data. Therefore, there is an urgent need for methods to automatically identify disturbance data segments within fault record data collected over a period of time. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is: how to analyze the waveform of the dataset collected by the fault recording device in the power network, and realize the location and identification of the disturbance dataset.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for identifying transient disturbance segments in fault recording data, comprising:

[0007] Collect fault waveform data;

[0008] Construct a dataset of waveform recording files and perform integrity checks on the dataset;

[0009] Filter and repair the waveform recording file dataset, and identify transient disturbances in the faulty waveform recording data;

[0010] A random forest model was constructed to calibrate the perturbation data segments in the fault recording data.

[0011] As a preferred embodiment of the method for identifying transient disturbance segments in fault recording data according to the present invention, the construction of the recording file dataset includes: if the recording file is missing a .dat data file, then the data set is deleted; if it is not missing, then the next step of judgment is performed; if the recording file is missing a .cfg configuration file, then the recording file of the same device is adaptively searched for and repaired; if it is not missing, then the storage and reading process is performed directly.

[0012] Select the dataset file storage path, and the machine will automatically read and load the waveform data in a loop; select the voltage and current channels according to the .cfg configuration file or adaptively, and read the channel identifier, sampling rate and absolute start time; read the waveform data in the .dat data file according to the selected voltage and current channels.

[0013] As a preferred embodiment of the method for identifying transient disturbance segments in fault recording data according to the present invention, the data filtering includes:

[0014] The sampling rate read determines the filter's structure parameter M, which is expressed as:

[0015] M = K / f

[0016] Where K is an empirical coefficient and f is the device sampling rate; filtering is performed channel by channel based on the determined structural parameters, and the typical filter formula is expressed as:

[0017]

[0018] Where, x (n) For the original signal, y (n) denoted as the filtered signal, and n as the data sequence.

[0019] As a preferred embodiment of the method for identifying transient disturbance segments in fault waveform data according to the present invention, the transient disturbance includes: taking the average difference between the original channel data and the filtered data; if the average difference result meets a threshold condition, then the fault waveform data of that channel is identified as having a transient disturbance; if it does not meet the threshold condition, then the waveform data of that channel is determined to have no disturbance, the calculation of this channel ends, and the calculation of the next channel begins; the threshold condition and the average difference of the signal are expressed as follows:

[0020] max{Del}>δ

[0021]

[0022] Where δ represents the disturbance detection threshold, Del (N) This indicates the signal uniformity.

[0023] As a preferred embodiment of the method for identifying transient disturbance segments in fault recording data according to the present invention, the construction of the random forest model includes extracting relevant features, voltage, current, and frequency from the fault recording data, and calculating the statistical properties of the relevant features.

[0024] For each decision tree, a subset of features and a subset of samples are randomly selected from the preprocessed dataset.

[0025] A decision tree is constructed using selected features and samples. The selection of each split node is based on information gain and Gini impurity. The information gain formula is expressed as:

[0026]

[0027] Gini impurity is expressed as:

[0028]

[0029] Where IG represents information gain; I represents the entropy of dataset D; D p D represents the dataset of the parent node; j Let f represent the child node dataset after the j-th split; f represents the data feature; p i This represents the proportion of the i-th class;

[0030] The predictions from all decision trees are aggregated; for classification problems, the majority voting method is used to solve for the final result.

[0031] As a preferred embodiment of the method for identifying transient disturbance segments in fault waveform data according to the present invention, the fault segment includes: if the waveform file is determined to have no missing .cfg configuration file, then data segments before and after the disturbance are calibrated based on the transient disturbance identification result and the absolute start time; and the data containing the disturbance is identified and calibrated multiple times using the mean difference result, and the disturbance channel and the data disturbance time are stored; if the .cfg configuration file is missing, the data disturbance time is not stored; the disturbance time is expressed as:

[0032] t dis =t start ±Δt

[0033] Among them, t dis Represents the disturbance time, based on the absolute start time t. start Calculation; Δt represents the relative time offset; the formula for the fault data segment is:

[0034]

[0035] Among them, D seg Let Δt represent the set of fault segments. before Indicates the time allowed before the disturbance; Δt after The time allowed after the disturbance is indicated, and f represents the sampling rate of the device.

[0036] The time period formula for the set of fault segments is expressed as:

[0037] T = [t] dis -Δt before ,t dis -Δt after ]

[0038] Where T represents the actual time segment of the fault segment set.

[0039] As a preferred embodiment of the method for identifying transient disturbance segments in fault recording data according to the present invention, the step of calibrating and classifying the data includes, let X = {x1, x2, ..., x...} n} represents the feature set of the fault recording data, and Y represents the category of the perturbation segment; the decision function of the random forest is expressed as:

[0040]

[0041] Where M represents the number of decision trees; T m Let θ represent the decision function of the m-th decision tree; m The set of parameters representing the decision tree is automatically selected by the random forest algorithm;

[0042] The decision function of each decision tree can be represented as a set of conditional branches and leaf nodes. A single decision tree is represented as follows:

[0043]

[0044] Where k represents the number of leaf nodes in the decision tree; w i G represents the weight of the i-th leaf node; i (X) represents a function based on the input data features X, used to calculate the path to the leaf node; τ i represents the threshold of the decision tree; I represents the indicator function, which determines whether the condition has been met.

[0045] The decision function of the random forest algorithm is used to classify the identified perturbation fragments.

[0046] A system for identifying transient disturbance segments in fault recording data, characterized in that:

[0047] Data acquisition module: Acquires fault waveform data, constructs waveform file datasets, and reads and organizes fault waveform file datasets;

[0048] Data processing module: Performs integrity checks on the files in the waveform recording files, filters and repairs the waveform recording file dataset, including adaptive search and repair of missing .cfg configuration files;

[0049] Identification module: Identifies transient disturbances in the processed fault recording data and marks the disturbance data segments;

[0050] Random Forest Model Module: Based on the difference between the filtered data and the original data, it determines whether there is a transient disturbance by setting a threshold condition and classifies the labeled disturbed data segments.

[0051] A computer device includes: a memory and a processor; the memory stores a computer program, characterized in that: when the processor executes the computer program, it implements the steps of the method described in any one of the present invention.

[0052] A computer-readable storage medium having a computer program stored thereon, characterized in that: when the computer program is executed by a processor, it implements the steps of the method described in any one of the present invention.

[0053] The beneficial effects of the present invention are as follows: The method for identifying transient disturbance segments in fault recording data provided by the present invention utilizes prior information of the dataset to achieve adaptive channel selection and file completion, and adaptively selects filter structure parameters based on the sampling rate. It analyzes the disturbance waveform characteristics in the recording dataset, can locate and identify disturbance data segments, and provide data support for the load side. Attached Figure Description

[0054] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0055] Figure 1 This is an overall flowchart of a method for identifying transient disturbance segments in fault recording data, provided in the first embodiment of the present invention.

[0056] Figure 2 The flowchart illustrates a method for identifying transient disturbance segments in fault recording data, as provided in the first embodiment of the present invention.

[0057] Figure 3 The second embodiment of the present invention provides a method for identifying transient disturbance segments in fault recording data, which includes the original waveform of the entire segment of the recording data containing the disturbance.

[0058] Figure 4 The waveform diagram of the perturbation in the recorded waveform data provided in the second embodiment of the present invention is shown below;

[0059] Figure 5 This is a runtime diagram of a method for identifying transient disturbance segments in fault recording data at different repetition counts, provided in the second embodiment of the present invention.

[0060] Figure 6 The original waveform of the entire segment of waveform data without disturbance is provided as the basis for a method for identifying transient disturbance segments in fault waveform data according to the second embodiment of the present invention.

[0061] Figure 7 The second embodiment of the present invention provides a method for identifying transient disturbance segments in fault recording data, which includes a waveform mean difference diagram of the entire segment of recording data without disturbance. Detailed Implementation

[0062] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0063] Example 1

[0064] Reference Figures 1-2 As an embodiment of the present invention, a method for identifying transient disturbance segments in fault recording data is provided, comprising:

[0065] S1: Construct a waveform recording dataset and perform integrity checks on the dataset.

[0066] Furthermore, the collected data includes: fault waveform data (including voltage and current waveforms, as well as corresponding frequency and phase information); equipment status data (operating status and historical performance data of equipment such as transformers, lines, and circuit breakers); environmental data (including weather conditions, temperature, and humidity); power grid structural information, such as line layout and connection methods; and historical fault data.

[0067] Furthermore, such as Figure 2 The dataset of waveform files collected by reading fault waveform data is shown; after reading, the integrity of the COMTRADE file in the dataset is checked.

[0068] Furthermore, after filtering and repairing the files, the waveform data file set is loaded; transient disturbances are identified and disturbance data segments are calibrated in the fault waveform data.

[0069] Furthermore, if the waveform recording file is missing a .dat data file, then that data set is deleted; otherwise, proceed to the next step of the judgment.

[0070] Furthermore, if the waveform recording file is missing the .cfg configuration file, it will adaptively search for waveform recording files from the same device and repair them; if it is not missing, it will proceed directly to the next step.

[0071] Furthermore, by selecting the dataset file storage path, the machine automatically reads and cyclically loads the waveform data; it selects the voltage and current channels according to the .cfg configuration file or adaptively, and reads the channel identifier, sampling rate, and absolute start-up time.

[0072] Furthermore, the waveform data in the .dat data file can be read according to the selected voltage channel and current channel.

[0073] It should be noted that data acquisition and preprocessing involves collecting multi-source data, such as fault waveforms, equipment status, and environmental data, and performing preprocessing. Data files are automatically selected and loaded, and voltage and current data are read and processed according to configuration files. Comprehensive data integrity checks and repairs ensure the accuracy and reliability of the dataset. This automated data processing and loading process significantly improves processing efficiency and reduces human error.

[0074] S2: Filter and repair the waveform recording file dataset, and identify transient disturbances in faulty waveform recording data.

[0075] Furthermore, constructing the waveform recording file dataset includes: if the waveform recording file is missing a .dat data file, then the data set is deleted; if it is not missing, then the next step of judgment is performed; if the waveform recording file is missing a .cfg configuration file, then the waveform recording file of the same device is adaptively searched for and repaired; if it is not missing, then the storage and reading process is performed directly.

[0076] Furthermore, the system selects the dataset file storage path, and the machine automatically reads and loads the waveform data in a loop; it selects the voltage and current channels according to the .cfg configuration file or adaptively, and reads the channel identifier, sampling rate, and absolute start time; it reads the waveform data from the .dat data file according to the selected voltage and current channels.

[0077] Furthermore, the sampling rate read determines the filter's structural parameters M, which are expressed as:

[0078] M = K / f

[0079] Where K is an empirical coefficient and f is the device sampling rate; filtering is performed channel by channel based on the determined structural parameters, and the typical filter formula is expressed as:

[0080]

[0081] Where, x (n) For the original signal, y (n) denoted as the filtered signal, and n as the data sequence.

[0082] Furthermore, transient disturbances include the average difference between the original channel data and the filtered data. If the average difference result meets the threshold condition, it is identified that there is a transient disturbance in the fault waveform data of that channel. If it does not meet the threshold condition, it is determined that there is no disturbance in the waveform data of that channel, the calculation of this channel ends, and the calculation of the next channel begins. The threshold condition and the average difference of the signal are expressed as follows.

[0083] max{Del}>δ

[0084]

[0085] Where δ represents the disturbance detection threshold, Del (N) This indicates the signal uniformity.

[0086] Furthermore, the fault segment includes, if the waveform recording file is not missing the .cfg configuration file, then based on the transient disturbance identification results and the absolute start time, calibrating the data segments before and after the disturbance; and using the mean difference results to repeatedly identify and calibrate the data containing the disturbance, storing the disturbance channel and the data disturbance time; if the .cfg configuration file is missing, then the data disturbance time is not stored; the disturbance time is represented as:

[0087] t dis =t start ±Δt

[0088] Among them, t dis Represents the disturbance time, based on the absolute start time t. start Calculation; Δt represents the relative time offset; the formula for the fault data segment is:

[0089]

[0090] Among them, D seg Let Δt represent the set of fault segments. before Indicates the time allowed before the disturbance; Δt after The value represents the time reserved after the disturbance, and f represents the sampling rate of the device.

[0091] Furthermore, the time period formula for the set of fault segments is expressed as:

[0092] T = [t] dis -Δt before ,t dis -Δt after ]

[0093] Where T represents the actual time segment of the fault segment set.

[0094] It should be noted that the process involves screening and repairing waveform recording datasets, automatically reading and loading waveform data, selecting voltage and current channels based on the .cfg file, reading the sampling rate and start-up time, and filtering the raw signal. Transient disturbances are identified by calculating the mean difference between the raw and filtered data and comparing it to a set threshold. Disturbance data segments are calibrated using the transient disturbance results and start-up time. This improves the automation of data processing, ensures data integrity and accuracy, accurately identifies transient disturbances through filtering and mean difference calculation, and makes fault analysis more precise and efficient.

[0095] S3: Construct a random forest model to calibrate the perturbation data segments in the fault recording data.

[0096] Furthermore, constructing a random forest model involves extracting relevant features from fault recording data, such as voltage, current, and frequency, and calculating the statistical properties of these features.

[0097] Furthermore, features and samples are randomly selected. For each decision tree, a subset of features and a subset of samples are randomly selected from the preprocessed dataset.

[0098] Furthermore, a decision tree is constructed using selected features and samples. The selection of each split node is based on information gain and Gini impurity. The information gain formula is expressed as:

[0099]

[0100] Furthermore, the impurity of the ginnie is expressed as:

[0101]

[0102] Where IG represents information gain; I represents the entropy of dataset D; D p D represents the dataset of the parent node; j Let f represent the child node dataset after the j-th split; f represents the data feature; p i This represents the proportion of the i-th class.

[0103] Furthermore, the predictions from all decision trees are aggregated; for classification problems, a majority voting method is used to solve for the final result.

[0104] Furthermore, let X = {x1, x2, ..., x} n} represents the feature set of the fault recording data, and Y represents the category of the perturbation segment; the decision function of the random forest is expressed as:

[0105]

[0106] Where M represents the number of decision trees; T mLet θ represent the decision function of the m-th decision tree; m The set of parameters representing the decision tree is automatically selected by the random forest algorithm.

[0107] Furthermore, the decision function of each decision tree can be represented as a set of conditional branches and leaf nodes, and a single decision tree is represented as:

[0108]

[0109] Where k represents the number of leaf nodes in the decision tree; w i G represents the weight of the i-th leaf node; i (X) represents a function based on the input data features X, used to calculate the path to the leaf node; τ i represents the threshold of the decision tree; I represents the indicator function, which judges whether the condition has been met.

[0110] Furthermore, the decision function of the random forest algorithm is used to classify the identified perturbation segments.

[0111] It should be noted that random forests improve the accuracy of classification and recognition of complex waveform data through ensemble learning of multiple decision trees. Randomly selected features and samples reduce the risk of overfitting and enhance the model's generalization ability on unseen data. Even if some decision trees perform poorly, the overall model can still maintain stable predictive performance. Suitable for large-scale data: Random forests are suitable for handling large datasets, and are particularly effective in data analysis of power systems.

[0112] On the other hand, this embodiment also provides a system for identifying transient disturbance segments in fault recording data, which includes:

[0113] Data acquisition module: Collects fault waveform data, constructs waveform file datasets, and reads and organizes fault waveform file datasets.

[0114] Data processing module: Performs integrity checks on the files in the waveform recording files, filters and repairs the waveform recording file dataset, including adaptive search and repair of missing .cfg configuration files.

[0115] Identification module: Identifies transient disturbances in the processed fault recording data and marks the disturbance data segments.

[0116] Random Forest Model Module: Based on the difference between the filtered data and the original data, it determines whether there is a transient disturbance by setting a threshold condition and classifies the labeled disturbed data segments.

[0117] If the above functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0118] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0119] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0120] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0121] Example 2

[0122] Reference Figure 3-7 As an embodiment of the present invention, a method for identifying transient disturbance segments in fault recording data is provided. To verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculations and simulation experiments.

[0123] Step 1: Determine the filter's structural parameters M based on the sampling rate. M can be calculated using the following formula:

[0124] M = K / f

[0125] In the formula, K is an empirical coefficient, which can be taken as 50000 in this paper, and f is the device sampling rate, which can be taken as 10000 in this paper;

[0126] Step 2: Filter by channel according to the structural parameters determined in Step 1. A typical filter is shown below:

[0127]

[0128] In the formula, x(n) is the original signal, y(n) is the filtered signal, and n is the data sequence. The mean difference between the original channel data and the filtered data is calculated. If the mean difference meets a threshold condition, a transient disturbance is identified in the recorded waveform data of that channel. If it does not meet the threshold condition, the recorded waveform data of that channel is determined to be free of disturbance, the calculation for this channel ends, and the calculation for the next channel begins. The threshold condition and the mean difference are given by the following two formulas:

[0129] max{Del}>δ

[0130]

[0131] In the formula, δ is the disturbance detection threshold, which can be 900 for text, and Del(n) is the signal mean difference;

[0132] like Figure 6 The waveform shown is the one without any disturbance. Figure 7As shown, the maximum mean difference of the waveform data segment without disturbance is 619, which is less than the disturbance detection threshold and is therefore judged as having no disturbance.

[0133] like Figure 3 The waveform shown is one with disturbance. Figure 5 The waveform of the extracted perturbation segment. Figure 4 As shown, the first wavefront value of the waveform mean difference of the waveform data segment with disturbance is 1875, which is greater than the disturbance detection threshold; therefore, it is determined that there is a disturbance.

[0134] Step 3: If Step 2 determines that the waveform recording file does not lack the .cfg configuration file, then based on the transient disturbance identification results and absolute start time of Step 2, calibrate the data segments before and after the disturbance, and use the mean difference results to identify and calibrate the data containing the disturbance multiple times, storing the disturbance channel and the data disturbance time; where the disturbance time is calculated as follows:

[0135] t dis =t start ±Δt

[0136] Where tdis is the disturbance time, calculated based on the absolute start time tstart = 14:26:40.714000s, and Δt is the relative time offset, which is 0.0041 in this example, so tdis = 40.7181. This gives the location of the disturbance time segment.

[0137] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for identifying transient disturbance segments in fault recording data, characterized in that, include: Collect fault waveform data; Construct a waveform recording dataset and perform integrity checks on the dataset; Filter and repair the waveform recording dataset, and identify transient disturbances in the faulty waveform recording data; A random forest model was constructed to calibrate the perturbation data segments in the fault recording data.

2. The method for identifying transient disturbance segments in fault recording data as described in claim 1, characterized in that: The construction of the waveform recording dataset includes: if the waveform recording file is missing a .dat data file, then the data set is deleted; if it is not missing, then the integrity of the cfg configuration file is checked. If the waveform recording file is missing the cfg configuration file, it will adaptively search for waveform recording files of the same device and repair them; if it is not missing, it will perform storage and reading processing. Select the dataset file storage path, and the machine will automatically read and load the waveform data in a loop; select the voltage and current channels according to the cfg configuration file or adaptively, and read the channel identifier, sampling rate and absolute start time; read the waveform data in the dat data file according to the selected voltage and current channels.

3. The method for identifying transient disturbance segments in fault recording data as described in claim 2, characterized in that: The data filtering includes, The sampling rate read determines the filter's structure parameter M, which is expressed as: M = K / f Where K represents an empirical coefficient and f represents the device sampling rate; filtering is performed channel by channel based on the determined structural parameters, and a typical filter formula is expressed as: Where, x (n) Represents the original signal, y (n) denoted as the filtered signal, and n represents the data sequence.

4. The method for identifying transient disturbance segments in fault recording data as described in claim 3, characterized in that: The transient disturbance includes the average difference between the original channel data and the filtered data. When the average difference result meets the threshold condition, the transient disturbance in the fault waveform data of the channel is identified. If it does not meet the threshold condition, the waveform data of the channel is determined to be free of disturbance, the calculation of this channel ends, and the calculation of the next channel begins. The threshold condition and the average difference of the signal are expressed as follows: max{Del}>δ Where δ represents the disturbance detection threshold, Del (N) This indicates the signal uniformity.

5. The method for identifying transient disturbance segments in fault recording data as described in claim 4, characterized in that: The construction of the random forest model includes extracting relevant features such as voltage, current, and frequency from fault recording data, and calculating the statistical properties of the relevant features. For each decision tree, a subset of features and a subset of samples are randomly selected from the preprocessed dataset. A decision tree is constructed using selected features and samples. The selection of each split node is based on information gain and Gini impurity. The information gain formula is expressed as: Gini impurity is expressed as: Where IG represents information gain; I represents the entropy of dataset D; and D... p D represents the dataset of the parent node; j Let f represent the child node dataset after the j-th split; f represents the data feature; p i This represents the proportion of the i-th class; the predictions from all decision trees are aggregated; for classification problems, the majority voting method is used to solve for the final result.

6. The method for identifying transient disturbance segments in fault recording data as described in claim 5, characterized in that: The fault segment includes, if the waveform recording file is not missing the cfg configuration file, then based on the transient disturbance identification results and the absolute start time, calibrating the data segments before and after the disturbance; and using the mean difference result to repeatedly identify and calibrate the data containing the disturbance, storing the disturbance channel and the data disturbance time; if the cfg configuration file is missing, then the data disturbance time is not stored; the disturbance time is represented as: t dis =t start ±Δt Among them, t dis Represents the disturbance time, based on the absolute start time t. start Calculation; Δt represents the relative time offset; the formula for the fault data segment is: Among them, D seg Let Δt represent the set of fault segments. before Indicates the time allowed before the disturbance; Δt after The time allowed after the disturbance is indicated, and f represents the sampling rate of the device. The time period formula for the set of fault segments is expressed as: T=[t dis -Δt before ,t dis -Δt after ] Where T represents the actual time segment of the fault segment set.

7. The method for identifying transient disturbance segments in fault recording data as described in claim 6, characterized in that: The disturbance data segments in the calibration fault recording data include Let X = {x1, x2, ..., x} n } represents the feature set of the fault recording data, and Y represents the category of the perturbation segment; the decision function of the random forest is expressed as: Where M represents the number of decision trees; T m Let θ represent the decision function of the m-th decision tree; m The set of parameters representing the decision tree is automatically selected by the random forest algorithm; The decision function of each decision tree can be represented as a set of conditional branches and leaf nodes. A single decision tree is represented as follows: Where k represents the number of leaf nodes in the decision tree; w i G represents the weight of the i-th leaf node; i (X) represents a function based on the input data features X, used to calculate the path to the leaf node; τ i represents the threshold of the decision tree; I represents the indicator function, which judges whether the condition is met; the decision tree identifies the types of perturbation segments, and the decision function of the random forest algorithm is used to classify and label the identified perturbation segments.

8. A system for identifying transient disturbance segments in fault recording data using the method described in any one of claims 1-7, characterized in that: Data acquisition module: Collects fault waveform data, constructs waveform file datasets, and reads and organizes fault waveform file datasets; Data processing module: Performs integrity checks on the files in the waveform recording files, filters and repairs the waveform recording file dataset, including adaptive search and repair of missing cfg configuration files; Identification module: Identifies transient disturbances in the processed fault recording data and marks the disturbance data segments; Random Forest Model Module: Based on the difference between the filtered data and the original data, it determines whether there is a transient disturbance by setting a threshold condition and classifies the labeled disturbed data segments.

9. A computer device, comprising: A memory and a processor; the memory stores a computer program, characterized in that: when the processor executes the computer program, it implements the steps of a method for identifying transient disturbance segments in fault recording data.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the steps of a method for identifying transient disturbance segments in fault recording data.

Images