Cable transient voltage identification method based on multi-domain features and adaptive Bayesian

Abstract

The invention relates to the technical field of cable transient voltage identification, and particularly provides a cable transient voltage identification method based on multi-domain features and adaptive Bayesian. The method comprises the following steps: acquiring a cable transient voltage, and preprocessing the cable transient voltage to obtain a preprocessed cable transient voltage; extracting multi-domain features of the preprocessed transient voltage of the cable to obtain comprehensive features; classification identification and feedback adjustment are carried out on the transient voltage of the cable through the comprehensive characteristics so as to update an identification strategy; according to the method, the problems of mode confusion and low algorithm robustness caused by insufficient sensitivity of a single feature are solved, and the accuracy and adaptability of cable transient voltage identification are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of cable transient voltage identification, and particularly to a cable transient voltage identification method based on multi-domain features and adaptive Bayesian. Background Technique

[0002] In complex system scenarios such as urban power grid connection, due to the characteristics of cable lines such as wide coverage and many electrical coupling paths, not only are they vulnerable to the influence of operation impact transient voltage disturbances, but also when adjacent equipment suffers transient overvoltage shocks, conductive overvoltages will be generated through electromagnetic coupling or grounding circuits, thus triggering chain insulation risks. The high amplitude, short duration, and non-linear waveform characteristics of transient overvoltages cause huge pressure on the cable insulation layer. Under such conditions for a long time, the insulation performance of the cable will gradually deteriorate due to the cumulative damage of transient overvoltages, and even local discharge defects will occur. Cables will be damaged to varying degrees under the action of different types of transient overvoltages (such as switching overvoltages, lightning overvoltages, short-circuit faults). Therefore, accurate identification of the types of cable transient overvoltages will provide an effective evaluation basis for the decline of the insulation state of cables due to cumulative damage caused by transient overvoltages, thus ensuring the safe and stable operation of cables and power systems.

[0003] Regarding the transient voltage type identification method proposed in the publicly disclosed patent document CN202010478764.7, during the mathematical morphology decomposition and wavelet transform processes, there may be insufficient adaptability to data. When the operating conditions of the offshore wind farm change significantly, it may lead to a decrease in the accuracy of decomposition and transformation, affecting the accuracy of transient voltage type identification. In addition, the energy value calculation and multi-level classification and identification links may face computational efficiency issues. In a complex offshore wind farm environment, a large amount of data processing requirements may result in a long response time for this method in practical applications, unable to meet the requirements of transient voltage monitoring scenarios with high real-time requirements. For the transient voltage type identification method and system proposed in the patent document CN202211528236.3, although it uses accurate real-time current signals to replace the distorted transient voltage signals on the secondary side of the CVT or the transient voltage measurement technology based on the voltage division principle to complete overvoltage type identification, improving the accuracy of identification, there may still be some deficiencies. This system has a high dependence on current signal acquisition devices. If devices such as current transformers experience a decrease in accuracy or failure, it may affect the accuracy of the entire identification system. In addition, in terms of feature quantity extraction and type identification algorithms, when facing complex power grid interference, misjudgment may occur, especially when there are multiple interferences superimposed in the power grid. Relying solely on the feature quantities of high-voltage capacitive current signals may not be able to accurately distinguish different types of transient voltages. For the voltage sag source identification method based on the BAS-BP classifier model proposed in the patent document CN202011416089.1, although it improves the correct rate and classification effect of voltage sag source identification by means of improving the S transform, combined weighting method, and beetle antennae search algorithm, there may still be some deficiencies. The process of constructing the feature index system and screening indicators in this method is relatively complex and requires a large amount of data processing and analysis work, which may result in a long time-consuming in the early feature processing stage in practical applications, especially when the data volume is large, affecting the efficiency of the entire identification system. In addition, although the BP neural network has been improved, there may still be limitations when dealing with voltage sag source identification problems with highly nonlinear or complex correlation relationships, resulting in an unsatisfactory classification effect. For the voltage sag classification method based on the integrated classifier proposed in the patent document CN202210688031.5, during the construction of the training sample set and the training process of the basic classifier, the training effect of the basic classifier may be poor due to unreasonable initial weight assignment. Especially when the sample quantity is unbalanced, the features of minority class samples are easily ignored, affecting the identification ability of the classifier for different sag types. In addition, although the weight update rule can adjust according to the sample quantity, in a complex power grid environment, the fixed update rule is difficult to adapt to the changes in sample feature distribution and importance, resulting in poor optimization effect of the classifier. These problems need to be solved in practical applications to improve the reliability and practicality of the method.

[0004] The existing cable transient voltage identification technology has significant defects such as single feature extraction dimension and insufficient algorithm adaptability: First, traditional methods mostly rely on single-dimension features such as time-domain amplitude and frequency-domain harmonic content, and fail to effectively fuse multi-domain features, resulting in a high pattern confusion rate under complex working conditions (such as the superposition of multiple types of overvoltages and noise interference). The amplitude ranges of lightning overvoltage and switching overvoltage may overlap, and false judgments are easily caused by relying solely on threshold criteria. Second, mainstream classification algorithms (support vector machines, decision trees, etc.) have problems such as low calculation efficiency, weak model interpretability, and difficulty in dealing with non-linear correlations between features. Summary of the Invention

[0005] In view of this, the present invention provides a cable transient voltage identification method based on multi-domain features and adaptive Bayesian to solve the problems of pattern confusion caused by insufficient sensitivity of single features and low algorithm robustness, and improve the accuracy and adaptability of cable transient voltage identification.

[0006] In a first aspect, the present invention provides a cable transient voltage identification method based on multi-domain features and adaptive Bayesian, and the method includes: Step 1: Collect the cable transient voltage and preprocess it to obtain the preprocessed cable transient voltage; Step 2: Extract multi-domain features from the preprocessed cable transient voltage to obtain comprehensive features; Step 3: Classify and identify the cable transient voltage through the comprehensive features and perform feedback adjustment to update the identification strategy; Step 4: Output the identification result according to the updated identification strategy.

[0007] Optionally, the collection of the cable transient voltage signal in Step 1 includes: First, obtain data through simulation: Build a power system simulation model. Use ATP-EMTP circuit simulation to construct a power system simulation model. The power system simulation model includes: a transmission line module, select a distributed parameter model, and set the overhead-cable line length and voltage level; a transformer module, select a transformer module with the function of studying transient overvoltage characteristics, and set relevant parameters such as transformer capacity and voltage ratio; a circuit breaker module, simulate real operation characteristics, and set the action time and switching curves of the circuit breaker; According to the sources of different transient overvoltages, set typical working conditions for simulation: (1) Voltage formula under switching operation conditions: The transient voltage caused by switching operation introduces variables related to multi-domain features and adaptive Bayesian, and its expression is: ; Among them, is the voltage mutation caused by switching operation; is the voltage amplitude at the moment of operation; is the voltage decay time constant, which is determined by the system impedance and capacitance parameters; is the angular frequency, which is used to reflect the frequency characteristics of the switching overvoltage; is the initial phase angle; is the operation uncertainty factor, which is an adaptive adjustment based on Bayesian; is the operation delay time; is the change in system impedance; (2) Voltage formula under lightning strike condition: The transient voltage caused by lightning strike introduces variables related to multi-domain characteristics and adaptive Bayesian, and its expression is: ; Wherein, is the change of the transient voltage caused by lightning strike over time; is the peak value of the lightning strike voltage; are the waveform rise and fall time constants respectively; is the waveform distortion coefficient, which is used to reflect the non-linear characteristics of the lightning strike waveform; is the frequency characteristic, which is used to reflect the frequency spectrum distribution of the lightning strike voltage; is the probability distribution of the lightning strike position, which is an adaptive adjustment based on Bayesian; (3) Voltage formula under short-circuit fault condition: The transient voltage caused by the short-circuit fault introduces variables related to multi-domain characteristics and adaptive Bayesian, and its expression is: ; Wherein, is the transient voltage caused by the short-circuit fault; is the system operating voltage; is the fault impedance; is the system impedance; is the fault voltage attenuation coefficient; is the fault voltage decay time constant; is the probability distribution of the fault position, which is an adaptive adjustment based on Bayesian; Secondly, data is acquired through the acquisition device: The acquisition device includes a measuring device for transient voltage and a collection unit for high-frequency signals; The measuring device for transient voltage is used to capture transient voltage signals, and the collection unit for high-frequency signals is used to collect high-frequency signals.

[0008] Optionally, the preprocessing of the cable transient voltage in step 1 to obtain the preprocessed cable transient voltage includes: Preprocess the traveling wave signal of the collected cable transient voltage, including removing low-frequency components and background noise, clearing the noise through filtering technology, so as to highlight the key features of the signal and obtain the preprocessed cable transient voltage.

[0009] Optionally, step 2 includes: Multi-domain feature extraction includes time-domain feature extraction, frequency-domain feature extraction, and time-frequency domain feature extraction, which are used to extract features of the cable transient voltage from different dimensions. By constructing a time-domain - frequency-domain - time-frequency domain multi-dimensional feature system, complementary features such as peak voltage, oscillation duration, proportion of fundamental frequency components, energy integral of a specific frequency band, short-time Fourier transform (STFT) time-frequency matrix, and wavelet coefficient energy distribution are extracted; The expression for time-domain feature extraction is: , where is the peak voltage, is the oscillation duration; The expression for frequency-domain feature extraction is: , where is the proportion of fundamental frequency components, is the energy integral of a specific frequency band; The expression for time-frequency domain feature extraction is: , where is the short-time Fourier transform (STFT) time-frequency matrix, is the wavelet coefficient energy distribution; Integrate the time-domain, frequency-domain, and time-frequency domain features into a comprehensive feature , and its expression is: .

[0010] Optionally, step 3 includes: Introduce a feature weight dynamic adjustment mechanism and prior probability online update; feature weight dynamic adjustment is used for feature optimization, and its expression is: ; where is the classification probability of the cable transient voltage; are the weight parameters of the adaptive Bayesian model respectively, which are optimized through training data; Prior probability online update is used for parameter adjustment, and its expression is: ; where is the prior probability at time ; is the prior probability at time ; is the smoothing factor, and its value ranges from 0.8 to 0.99; is the time The number of samples belonging to the current category; is the time of the total number of samples; Combining feature weight dynamic adjustment and prior probability online update into an adaptive Bayesian recognition model to update the recognition strategy, and its expression is: ; where, is the final classification probability; is the updated prior probability, that is, the prior probability at time ; is the weighted sum of the prior probabilities and classification probabilities of all categories.

[0011] Optionally, the step 4 includes: a. Classification label: According to the output of the adaptive Bayesian model, identify the specific type of transient overvoltage, where the label definitions are as follows: Category = 1, indicating switching overvoltage; Category = 2, indicating lightning overvoltage; Category = 3, indicating short-circuit fault; b. Result output: Map the classification label to the corresponding transient overvoltage type to form the final recognition result.

[0012] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium, where the computer-readable storage medium includes a stored program, and when the program runs, it controls the device where the computer-readable storage medium is located to execute the cable transient voltage recognition method based on multi-domain features and adaptive Bayesian in the first aspect or any possible implementation manner of the first aspect.

[0013] In a fourth aspect, an embodiment of the present invention provides an electronic device, including: one or more processors; a memory; and one or more computer programs, where the one or more computer programs are stored in the memory, and the one or more computer programs include instructions, and when the instructions are executed by the device, the device executes the cable transient voltage recognition method based on multi-domain features and adaptive Bayesian in the first aspect or any possible implementation manner of the first aspect.

[0014] In the technical solution provided by the present invention, the method includes collecting the cable transient voltage, preprocessing it to obtain the preprocessed cable transient voltage; extracting multi-domain features of the preprocessed cable transient voltage to obtain comprehensive features; classifying and recognizing the cable transient voltage through the comprehensive features and performing feedback adjustment to update the recognition strategy; and outputting the recognition result according to the updated recognition strategy. This method solves the problems of pattern confusion caused by insufficient sensitivity of single features and low algorithm robustness, and improves the accuracy and adaptability of cable transient voltage recognition. Description of the Drawings

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0016] Figure 1 It is a flowchart of the cable transient voltage identification method based on multi-domain features and adaptive Bayesian provided by the embodiments of the present invention; Figure 2 It is a schematic diagram of multi-domain feature extraction provided by the embodiments of the present invention; Figure 3 It is a schematic diagram of updating the identification strategy provided by the embodiments of the present invention; Figure 4 It is a schematic diagram of an electronic device provided by the embodiments of the present invention. Detailed implementation manners

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0018] It should be clear that the described embodiments are only some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0019] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The singular forms "a", "the", and "said" used in the embodiments of the present invention are also intended to include the plural forms unless the context clearly indicates otherwise.

[0020] It should be understood that the term " / and / " used herein is only a description of the association relationship of associated objects, indicating that three relationships may exist. For example, a / or b may represent: a exists alone, a and b exist simultaneously, and b exists alone. In addition, the character " / " herein generally represents an "or" relationship between the preceding and following associated objects.

[0021] Depending on the context, as used herein, the term "if" can be interpreted as "when" or "while" or "in response to determining" or "in response to detecting". Similarly, depending on the context, the phrase "if determined" or "if detecting (stated condition or event)" can be interpreted as "when determined" or "in response to determining" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)".

[0022] With the improvement of the cable rate in the power system and the high - proportion access of new energy, due to characteristics such as wide distribution area and many electrical coupling paths of cable lines, they are vulnerable to transient over - voltages. Such over - voltages have characteristics such as high amplitude, non - linear waveform and short action time. Long - term action will lead to a decline in the insulation state due to the cumulative damage of transient over - voltages, resulting in cable deterioration, and even local discharge defects, significantly threatening the reliability of power grid operation. The types of transient over - voltages are mixed, and traditional identification methods are difficult to meet the accurate classification requirements.

[0023] Multi - domain feature extraction technology can capture the characteristics of data more comprehensively by fusing multi - dimensional features such as time domain, frequency domain and time - frequency domain, thus significantly improving the accuracy and robustness of identification and classification. This technology has been widely applied in many fields. In communication signal processing, multi - domain feature extraction is used for the identification and classification of radar signals, successfully improving the signal recognition accuracy in complex environments. In the power system, this technology is widely used in power quality monitoring, effectively improving the accuracy of fault diagnosis. In addition, in the field of industrial automation, multi - domain feature extraction technology is used for vibration signal analysis, successfully realizing early warning and diagnosis of equipment failures.

[0024] Adaptive Bayesian classification technology shows significant advantages in data classification and prediction tasks in complex environments by dynamically adjusting feature weights and prior probabilities. This technology can effectively handle the non - linear correlation between features and adapt to the changes in data distribution, thus improving the accuracy and adaptability of classification. In the field of medical diagnosis, adaptive Bayesian classification is used for disease prediction and diagnosis. By updating the prior probability in real - time to adapt to new diagnostic data, it significantly improves the accuracy and adaptability of diagnosis. In financial risk prediction, this technology adapts to market changes by dynamically adjusting feature weights, successfully improving the reliability of risk prediction. In addition, in the intelligent transportation system, adaptive Bayesian classification is used for traffic flow prediction and accident warning. By updating the prior probability in real - time to adapt to the changes in traffic flow, it significantly improves the accuracy of prediction and the robustness of the system.

[0025] Figure 1 The flowchart of the cable transient voltage identification method based on multi - domain features and adaptive Bayesian provided by the embodiments of the present invention is as Figure 1 shown, and the method includes: Step 1: Collect the transient voltage of the cable and preprocess it to obtain the preprocessed transient voltage of the cable.

[0026] In the embodiment of the present invention, the collection of the transient voltage signal of the cable in Step 1 includes: First, obtain data by simulation: Build a power system simulation model. Use ATP-EMTP circuit simulation to construct a power system simulation model. The power system simulation model includes: a transmission line module, select a distributed parameter model, and set the length and voltage level of the overhead-cable line; a transformer module, select a transformer module with the function of studying transient overvoltage characteristics, and set relevant parameters such as the transformer capacity and voltage ratio; a circuit breaker module, simulate the real operation characteristics, and set the action time and switching curves of the circuit breaker; According to the sources of different transient overvoltages, set typical working conditions for simulation: (1) Voltage formula under switching operation conditions: The transient voltage caused by switching operation introduces multi-domain characteristics (such as operation delay time, overvoltage frequency characteristics) and variables related to adaptive Bayesian, and its expression is: ; Among them, is the voltage mutation caused by switching operation; is the voltage amplitude at the moment of operation; is the voltage decay time constant, which is determined by the system impedance and capacitance parameters; is the angular frequency, which is used to reflect the frequency characteristics of the switching overvoltage; is the initial phase angle; is the operation uncertainty factor, which is an adaptive adjustment based on Bayesian; is the operation delay time; is the change in system impedance; (2) Voltage formula under lightning strike conditions: The transient voltage caused by lightning strike introduces multi-domain characteristics (such as frequency characteristics, waveform distortion coefficient) and variables related to adaptive Bayesian, and its expression is: ; Among them, is the change of the transient voltage caused by lightning strike over time; is the peak value of the lightning strike voltage; are the waveform rise and fall time constants respectively; is the waveform distortion coefficient, which is used to reflect the non-linear characteristics of the lightning strike waveform; is the frequency characteristic, which is used to reflect the frequency spectrum distribution of the lightning strike voltage; is the probability distribution of the lightning strike position, which is an adaptive adjustment based on Bayesian; (3) Voltage formula under short-circuit fault conditions: The transient voltage caused by a short - circuit fault introduces multi - domain features (such as the uncertainty of fault impedance and the probability distribution of fault location) and variables related to adaptive Bayesian, and its expression is: ; Wherein, is the transient voltage caused by a short - circuit fault; is the system operating voltage; is the fault impedance; is the system impedance; is the fault voltage attenuation coefficient; is the fault voltage attenuation time constant; is the probability distribution of the fault location, which is an adaptive adjustment based on Bayesian; Secondly, data is acquired through an acquisition device: The acquisition device includes a measuring device for high - precision transient voltage and a collection unit for high - frequency signals; The measuring device for transient voltage is used to capture high - precision transient voltage signals, and the parameters include sampling frequency, voltage range, resolution, etc.; The collection unit for high - frequency signals (such as lightning strikes, switch operations) is used to collect high - frequency signals, and the parameters include bandwidth, sampling rate, anti - aliasing filter, etc.

[0027] In the embodiment of the present invention, in step 1, the cable transient voltage is pre - processed, and the pre - processed cable transient voltage obtained includes: The traveling - wave signal of the collected cable transient voltage is pre - processed, including removing low - frequency components and background noise to clear the noise through filtering technology so as to highlight the key features of the signal and obtain the pre - processed cable transient voltage.

[0028] Step 2: Extract multi - domain features from the pre - processed cable transient voltage to obtain comprehensive features.

[0029] In the embodiment of the present invention, as Figure 2 shown, step 2 includes: Multi - domain feature extraction includes time - domain feature extraction, frequency - domain feature extraction, and time - frequency domain feature extraction, which are used to extract features of the cable transient voltage from different dimensions. By constructing a time - domain - frequency - domain - time - frequency domain multi - dimensional feature system, complementary features such as peak voltage, oscillation duration, proportion of fundamental frequency component, energy integral of a specific frequency band, short - time Fourier transform (STFT) time - frequency matrix, and wavelet coefficient energy distribution are extracted; This method solves the problem of pattern confusion caused by insufficient sensitivity of single features, improves the feature representation ability, and provides more comprehensive and accurate feature data for subsequent classification and recognition.

[0030] The expression of time - domain feature extraction is: , wherein, is the peak voltage, is the oscillation duration; The expression for frequency-domain feature extraction is: , where is the proportion of the fundamental frequency component, is the energy integration of a specific frequency band (such as kHz - MHz); The expression for time-frequency domain feature extraction is: , where is the short-time Fourier transform (STFT) time-frequency matrix, is the energy distribution of wavelet coefficients; Integrate the time-domain, frequency-domain, and time-frequency domain features into a comprehensive feature , and its expression is: .

[0031] Step 3: Through the comprehensive feature, classify and identify the transient voltage of the cable and perform feedback adjustment to update the identification strategy.

[0032] In the embodiment of the present invention, as Figure 3 shown, Step 3 includes: Updating the identification strategy includes dynamic adjustment of feature weights, an adaptive Bayesian identification model, i.e., an adaptive Bayesian classifier, online update of prior probabilities, feature optimization, parameter adjustment, and the identified result of the output; introducing a dynamic adjustment mechanism for feature weights and an online update of prior probabilities; the dynamic adjustment of feature weights is used for feature optimization, and its expression is: ; where is the classification probability of the transient voltage of the cable; are the weight parameters of the adaptive Bayesian model respectively, which are optimized through training data; The online update of prior probabilities is used for parameter adjustment, and its expression is: ; where is the prior probability at time ; is the prior probability at time ; is the smoothing factor, and its value ranges from 0.8 to 0.99; is the number of samples belonging to the current category at time ; is the total number of samples at time ; Combine the dynamic adjustment of feature weights and the online update of prior probabilities into the adaptive Bayesian identification model to update the identification strategy, and its expression is: ; where is the final classification probability; is the updated prior probability, i.e., the prior probability at time of the prior probability; is the weighted sum of the prior probabilities and classification probabilities of all categories.

[0033] In the embodiments of the present invention, the above expression summarizes the mechanism of dynamic adjustment of feature weights and online update of prior probabilities, ensuring that the adaptive Bayesian recognition model can adapt to changes in data distribution and improve the classification accuracy.

[0034] Step 4: Output the recognition result according to the updated recognition strategy.

[0035] In the embodiments of the present invention, Step 4 includes: a. Classification label: According to the output of the adaptive Bayesian model, identify the specific type of transient overvoltage, where the label definitions are as follows: Category = 1 indicates switching overvoltage; Category = 2 indicates lightning overvoltage; Category = 3 indicates short - circuit fault; b. Result output: Map the classification label to the corresponding transient overvoltage type to form the final recognition result. For example, if the input feature vector is , and its classification label is: Category = 2, then the recognition result is lightning overvoltage.

[0036] In the embodiments of the present invention, for the convenience of analysis, the recognition result is presented in the form of a table or a graph. When presented in tabular form, it is shown in Table 1.

[0037] Table 1 .

[0038] The present invention constructs a multi-dimensional feature system in the time domain, frequency domain, and time-frequency domain, extracts complementary features such as the peak voltage of the transient voltage, oscillation duration, peak-to-peak ratio, energy integral in a specific frequency band (such as kHz - MHz), energy distribution of wavelet coefficients, and time-frequency matrix of short-time Fourier transform (STFT), effectively improving the feature representation ability, and solving the problem of pattern confusion caused by insufficient sensitivity of traditional methods in single features. Existing technologies mostly rely on single-dimensional features such as time-domain amplitude or frequency-domain harmonic content, which are difficult to meet the requirements of high-precision identification. By introducing a dynamic adjustment mechanism of feature weights and combining it with online update of prior probability into an adaptive Bayesian recognition model, the present invention designs an adaptive Bayesian classifier, weakens the constraint of the feature independence assumption, and improves the classification accuracy of the algorithm in complex scenarios such as noise interference and data distribution offset. This dynamic adjustment mechanism enables the classifier to be optimized in real time, ensuring high-precision classification performance in a changing power grid operation environment, effectively improving the intelligent level of power system transient voltage identification. In contrast, the classification algorithms in existing technologies generally lack a dynamic adaptation mechanism and are difficult to handle the problem of feature distribution offset caused by power grid topology changes, equipment aging, etc. The classifier of the present invention can maintain a high classification accuracy in complex scenarios, effectively improving the robustness and adaptability of the algorithm; through the above technical innovations, the present invention significantly improves the accuracy and adaptability of cable transient voltage identification, providing core technical support for power system transient process analysis and active defense.

[0039] In the technical solution provided by the present invention, the method includes collecting the cable transient voltage, preprocessing it to obtain the preprocessed cable transient voltage; extracting multi-domain features of the preprocessed cable transient voltage to obtain comprehensive features; classifying and identifying the cable transient voltage through the comprehensive features and performing feedback adjustment to update the identification strategy; and outputting the identification result according to the updated identification strategy. This method solves the problems of pattern confusion caused by insufficient sensitivity of single features and low algorithm robustness, and improves the accuracy and adaptability of cable transient voltage identification.

[0040] Each step of the embodiment of the present invention can be executed by an electronic device. Among them, the electronic device includes but is not limited to mobile phones, tablet computers, portable PCs, desktop computers, etc.

[0041] The embodiment of the present invention provides a computer-readable storage medium. The computer-readable storage medium includes a stored program. When the program runs, it controls the electronic device where the computer-readable storage medium is located to execute the embodiment of the above-mentioned cable transient voltage identification method based on multi-domain features and adaptive Bayesian.

[0042] Figure 4 It is a schematic diagram of an electronic device provided by the embodiment of the present invention, as Figure 4As shown, the electronic device 21 includes: a processor 211, a memory 212, and a computer program 213 stored in the memory 212 and executable on the processor 211. When the computer program 213 is executed by the processor 211, it implements the cable transient voltage identification method based on multi-domain features and adaptive Bayesian in the embodiments. To avoid repetition, details are not described here one by one.

[0043] The electronic device 21 includes, but is not limited to, a processor 211 and a memory 212. Those skilled in the art can understand that Figure 4 These are merely examples of the electronic device 21 and do not constitute a limitation on the electronic device 21. It may include more or fewer components than shown in the figure, or combine certain components, or have different components. For example, the electronic device may also include input / output devices, network access devices, buses, etc.

[0044] The so-called processor 211 may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0045] The memory 212 may be an internal storage unit of the electronic device 21, such as the hard disk or memory of the electronic device 21. The memory 212 may also be an external storage device of the electronic device 21, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., equipped on the electronic device 21. Further, the memory 212 may also include both the internal storage unit and the external storage device of the electronic device 21. The memory 212 is used to store computer programs and other programs and data required by the network device. The memory 212 may also be used to temporarily store data that has been output or will be output.

[0046] Those skilled in the art can clearly understand that for the sake of convenience and brevity of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated here.

[0047] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A cable transient voltage identification method based on multi-domain features and adaptive Bayesian, characterized in that, The method includes the following steps: Step 1: Collect the transient voltage of the cable, and perform preprocessing on it to obtain the preprocessed transient voltage of the cable; Step 2: Extract multi-domain features from the preprocessed transient voltage of the cable to obtain comprehensive features; Step 3: Through the comprehensive features, classify and identify the transient voltage of the cable and perform feedback adjustment to update the identification strategy; Step 4: Output the identification result according to the updated identification strategy.

2. The method according to claim 1, wherein In step 1, collecting the transient voltage signal of the cable includes: First, obtain data by simulation: Build a power system simulation model, use ATP-EMTP circuit simulation to construct a power system simulation model. The power system simulation model includes: a transmission line module, select a distributed parameter model, and set the length of the overhead-cable line and the voltage level; a transformer module, select a transformer module with the function of studying transient overvoltage characteristics, and set relevant parameters such as the transformer capacity and voltage ratio; a circuit breaker module, simulate the real operation characteristics, and set the action time and switching curves of the circuit breaker; According to the sources of different transient overvoltages, set typical working conditions for simulation: (1) Voltage formula under switching operation conditions: The transient voltage caused by switching operation introduces variables related to multi-domain features and adaptive Bayesian, and its expression is: ； Among them, is the voltage mutation caused by the switch operation; is the voltage amplitude at the instant of operation; is the voltage decay time constant, which is determined by the system impedance and capacitance parameters; is the angular frequency, which is used to reflect the frequency characteristics of the switching overvoltage; is the initial phase angle; is the operation uncertainty factor, which is an adaptive adjustment based on Bayesian; is the operation delay time; is the change in system impedance; (2) Voltage formula under lightning strike conditions: The transient voltage caused by lightning strike introduces variables related to multi-domain features and adaptive Bayesian, and its expression is: ； Among them, is the transient voltage caused by lightning strike varying with time; is the peak value of the lightning strike voltage; are the waveform rise and fall time constants respectively; is the waveform distortion coefficient, which is used to reflect the non-linear characteristics of the lightning strike waveform; is the frequency characteristic, which is used to reflect the spectrum distribution of the lightning strike voltage; is the probability distribution of the lightning strike position, which is an adaptive adjustment based on Bayesian; (3) Voltage formula under short-circuit fault conditions: The transient voltage caused by short-circuit fault introduces variables related to multi-domain features and adaptive Bayesian, and its expression is: ； Among them, is the transient voltage caused by the short-circuit fault; is the system operating voltage; is the fault impedance; is the system impedance; is the fault voltage attenuation coefficient; is the fault voltage attenuation time constant; is the probability distribution of the fault location, which is adaptively adjusted based on Bayesian; Secondly, obtain data through a collection device: The collection device includes a measuring device for transient voltage and a collection unit for high-frequency signals; the measuring device for transient voltage is used to capture transient voltage signals, and the collection unit for high-frequency signals is used to collect high-frequency signals.

3. The method according to claim 1, wherein In step 1, preprocessing the transient voltage of the cable to obtain the preprocessed transient voltage of the cable includes: Preprocess the traveling wave signal of the collected transient voltage of the cable, including removing low-frequency components and background noise, and clearing the noise through filtering technology to highlight the key features of the signal, so as to obtain the preprocessed transient voltage of the cable.

4. The method according to claim 1, characterized in that, Step 2 includes: Multi-domain feature extraction includes time-domain feature extraction, frequency-domain feature extraction, and time-frequency domain feature extraction, which are used to extract features of the transient voltage of the cable from different dimensions. By constructing a time-frequency-time-frequency multi-dimensional feature system, complementary features such as peak voltage, oscillation duration, proportion of fundamental frequency components, energy integral of a specific frequency band, short-time Fourier transform STFT time-frequency matrix, and wavelet coefficient energy distribution are extracted; The expression for time-domain feature extraction is as follows: , where is the peak voltage, is the oscillation duration; The expression for frequency-domain feature extraction is as follows: , where is the proportion of the fundamental frequency component, is the energy integral of a specific frequency band; The expression for time-frequency domain feature extraction is as follows: , where is the time-frequency matrix of the short-time Fourier transform (STFT), is the energy distribution of wavelet coefficients; Integrate time-domain, frequency-domain, and time-frequency-domain features into a comprehensive feature , and its expression is: .

5. The method according to claim 1, characterized in that, Step 3 includes: Introduce a dynamic adjustment mechanism for feature weights and online update of prior probabilities; the dynamic adjustment of feature weights is used for feature optimization, and its expression is: ； Among them, is the classification probability of the cable transient voltage; are the weight parameters of the adaptive Bayesian model, which are optimized by training data respectively; The online update of prior probabilities is used for parameter adjustment, and its expression is: ； wherein, is the prior probability of time ; is the prior probability of time ; is a smoothing factor, and its value ranges from 0.8 to 0.99; is the number of samples of time belonging to the current category; is the total number of samples of time ; Combine the dynamic adjustment of feature weights and the online update of prior probabilities into an adaptive Bayesian identification model to update the identification strategy, and its expression is: ； Among them, is the final classification probability; is the updated prior probability, that is, the prior probability at time ; is the weighted sum of the prior probabilities and classification probabilities of all categories.

6. The method according to claim 1, wherein Step 4 includes: a. Classification label: Identify the specific type of transient overvoltage according to the output of the adaptive Bayesian model, where the label definitions are as follows: Category = 1 indicates switching overvoltage; Category = 2 indicates lightning overvoltage; Category = 3 indicates short-circuit fault. b. Result output: Map the classification label to the corresponding transient overvoltage type to form the final recognition result.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein when the program runs, it controls the device where the computer-readable storage medium is located to execute the cable transient voltage recognition method based on multi-domain features and adaptive Bayesian as described in any one of claims 1 to 6.

8. An electronic device, characterized in that, Comprising: One or more processors; A memory; And one or more computer programs, wherein the one or more computer programs are stored in the memory, and the one or more computer programs include instructions that, when executed by the device, cause the device to execute the cable transient voltage recognition method based on multi-domain features and adaptive Bayesian as described in any one of claims 1 to 6.

Images