Line fault detection method based on frequency response transfer function, medium and equipment

Abstract

The present application relates to the technical field of electric power, and discloses a line fault detection method based on a frequency response transfer function, a medium and equipment, comprising the following steps: determining the frequency response of each frequency by the frequency domain component of each frequency in the historical frequency domain distortion signal and the frequency domain component of each frequency in the experimental frequency domain restoration signal, and obtaining a plurality of numerator transfer coefficients and a plurality of denominator transfer coefficients according to the frequency response of each frequency and a preset target function, so as to construct the frequency response transfer function; and restoring the amplitude and phase of the current frequency domain distortion signal according to the amplitude attenuation coefficient and phase offset of the frequency domain transfer function corresponding to the constructed frequency response transfer function. The method can effectively eliminate the nonlinear influence of the mutual inductor on the amplitude and phase of the measured signal, accurately restore the real fault information, and thus improve the detection accuracy of the hidden fault of the overhead line.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric power, and in particular to a line fault detection method based on a frequency response transfer function, a medium and equipment. BACKGROUND

[0002] With the continuous development of the power system, the safety and reliability of the distribution overhead line as a key component of the power supply network are increasingly valued. However, in the actual operation of the distribution overhead line, it is influenced by external environment and line itself, resulting in frequent faults, such as high resistance grounding fault and hidden fault. The traditional fault detection method mainly relies on direct measurement and analysis of the line electrical quantity, but in a complex electromagnetic environment, the signal often appears nonlinear distortion, which makes it extremely difficult to accurately extract the fault characteristics, and further leads to low detection accuracy of hidden faults.

[0003] The mutual inductor as a commonly used signal acquisition device in overhead lines has nonlinear effects on the amplitude and phase of the measured signal due to its non-ideal characteristics, which distorts the fault signal and makes the detection system unable to accurately restore the real fault information, thereby affecting the accuracy of fault detection. SUMMARY

[0004] Therefore, it is necessary to propose a line fault detection method based on a frequency response transfer function, a medium and equipment to effectively eliminate the nonlinear effects of the mutual inductor on the amplitude and phase of the measured signal, accurately restore the real fault information, and improve the detection accuracy of hidden faults in overhead lines.

[0005] To achieve the above purpose, the present application provides a line fault detection method based on a frequency response transfer function in the first aspect, which comprises:

[0006] obtaining a current frequency domain distorted signal and a historical frequency domain distorted signal of the mutual inductor, and an experimental frequency domain restoration signal corresponding to the historical frequency domain distorted signal;

[0007] determining the frequency response of each frequency according to the frequency domain component of each frequency in the historical frequency domain distorted signal and the frequency domain component of each frequency in the experimental frequency domain restoration signal, and obtaining a plurality of numerator transfer coefficients and a plurality of denominator transfer coefficients according to the frequency response of each frequency and a preset target function, to construct a frequency response transfer function;

[0008] transforming the frequency response transfer function to obtain a frequency domain form function, and determining a target frequency domain restoration signal according to the amplitude decay coefficient and phase offset of the frequency domain form function, and the amplitude and phase of each frequency in the current frequency domain distorted signal;

[0009] perform inverse Fourier transform on the target frequency domain restoration signal to obtain a target restoration signal;

[0010] perform multi-feature extraction on the target restoration signal to obtain a signal feature vector, and input the signal feature vector into a preset machine learning model to obtain a fault classification detection result.

[0011] Optionally, the frequency response of each frequency is determined according to the frequency domain component of each frequency in the historical frequency domain distortion signal and the frequency domain component of each frequency in the experimental frequency domain restoration signal, and the plurality of numerator transfer coefficients and the plurality of denominator transfer coefficients are obtained according to the frequency response of each frequency and a preset target function to construct the frequency response transfer function, including:

[0012] the quotient value between the frequency domain component of the dth frequency in the historical frequency domain distortion signal and the frequency domain component of the dth frequency in the experimental frequency domain restoration signal is taken as the frequency response of the dth frequency; wherein d takes an integer greater than 0 in turn until d is equal to the total number of frequencies of the historical frequency domain distortion signal or the experimental frequency domain restoration signal, to obtain the frequency response of each frequency;

[0013] using the LM method, the plurality of numerator transfer coefficients and the plurality of denominator transfer coefficients in the preset target function are adjusted according to the frequency response of each frequency to make the function value of the preset target function reach a minimum under the condition of satisfying the preset target function, to obtain the plurality of numerator transfer coefficients and the plurality of denominator transfer coefficients;

[0014] the frequency response transfer function is constructed according to the plurality of numerator transfer coefficients and the plurality of denominator transfer coefficients.

[0015] Optionally, the expression of the preset target function is:

[0016] ;

[0017] wherein, ;

[0018] in the above formula, is the function value of the preset target function, is the total number of frequencies of the historical frequency domain distortion signal or the experimental frequency domain restoration signal, is the frequency response of the dth frequency, is the Laplacian corresponding to the dth frequency, is the m+1th numerator transfer coefficient, is the total number of numerator transfer coefficients or denominator transfer coefficients, is the m+1th denominator transfer coefficient, is an imaginary unit, is a circular constant, is the dth frequency.

[0019] Optionally, the expression of the frequency response transfer function is:

[0020] ;

[0021] wherein, ;

[0022] In the above formula, is the frequency response transfer function, is the Laplace operator, is the m+1th numerator transfer coefficient, is the total number of numerator transfer coefficients or denominator transfer coefficients, is the m+1th denominator transfer coefficient, is the imaginary unit, is the circular constant, is the frequency.

[0023] Optionally, the expression of the frequency domain form function is:

[0024] ;

[0025] wherein, is the frequency domain form function, is the amplitude attenuation coefficient of the frequency f, is the natural constant, is the imaginary unit, is the phase offset of the frequency f.

[0026] Optionally, the determining the target frequency domain restoration signal according to the amplitude attenuation coefficient and the phase offset of the frequency domain form function, and the amplitude and the phase of each frequency in the current frequency domain distortion signal comprises:

[0027] determining the target frequency domain restoration signal by using the formula ;

[0028] wherein, is the kth frequency domain component of the kth frequency in the target frequency domain restoration signal, is the kth frequency, is the amplitude of the kth frequency in the current frequency domain distortion signal, is the amplitude attenuation coefficient of the frequency f k , is the natural constant, is the imaginary unit, is the phase of the kth frequency in the current frequency domain distortion signal, is the phase offset of the frequency f kThe phase offset amount.

[0029] Optionally, the multi-feature extraction on the target restoration signal comprises:

[0030] Wavelet transform is performed on the target restoration signal to obtain a first time-frequency graph;

[0031] Short-time Fourier transform is performed on the target restoration signal to obtain a second time-frequency graph;

[0032] Multi-feature extraction is performed on the first time-frequency graph and the second time-frequency graph to obtain a spectral center frequency, a frequency bandwidth, a short-time energy and a time-domain waveform variation amount;

[0033] The spectral center frequency, the frequency bandwidth, the short-time energy and the time-domain waveform variation amount are taken as the signal feature vector.

[0034] Optionally, the obtaining of the current frequency domain distortion signal of the mutual inductor comprises:

[0035] The current distortion signal of the mutual inductor is obtained;

[0036] Fourier transform is performed on the current distortion signal to obtain an initial frequency domain distortion signal;

[0037] The maximum frequency is determined according to the experimental frequency domain restoration signal;

[0038] The cutoff frequency is determined according to the maximum frequency and a preset safety coefficient;

[0039] The initial frequency domain distortion signal is frequency-screened according to the cutoff frequency to obtain the current frequency domain distortion signal.

[0040] To achieve the above object, the application provides a line fault detection device based on a frequency response transfer function in a second aspect, and the device comprises:

[0041] An acquisition module is configured to acquire a current frequency domain distortion signal and a historical frequency domain distortion signal of a mutual inductor, and an experimental frequency domain restoration signal corresponding to the historical frequency domain distortion signal;

[0042] A determination and construction module is configured to determine a frequency response of each frequency according to a frequency domain component of each frequency in the historical frequency domain distortion signal and a frequency domain component of each frequency in the experimental frequency domain restoration signal, and obtain a plurality of numerator transfer coefficients and a plurality of denominator transfer coefficients according to the frequency response of each frequency and a preset target function, so as to construct a frequency response transfer function;

[0043] The conversion and determination module is configured to convert the frequency response transfer function to obtain a frequency domain function, and determine a target frequency domain restoration signal according to an amplitude attenuation coefficient and a phase offset of the frequency domain function and an amplitude and a phase of each frequency in the current frequency domain distortion signal.

[0044] The inverse transform module is configured to perform inverse Fourier transform on the target frequency domain restoration signal to obtain a target restoration signal.

[0045] The extraction and prediction module is configured to perform multi-feature extraction on the target restoration signal to obtain a signal feature vector, and input the signal feature vector into a preset machine learning model to obtain a fault classification detection result.

[0046] To achieve the above object, the present application provides a computer readable storage medium storing a computer program in a third aspect, the computer program is executed by a processor, so that the processor executes the line fault detection method based on the frequency response transfer function as any one of the first aspect.

[0047] To achieve the above object, the present application provides a computer device including a memory and a processor in a fourth aspect, the memory stores a computer program, the computer program is executed by the processor, so that the processor executes the line fault detection method based on the frequency response transfer function as any one of the first aspect.

[0048] The embodiment of the present application has the following beneficial effects: the method obtains the current frequency domain distortion signal and the historical frequency domain distortion signal of the mutual inductor, and the experimental frequency domain restoration signal corresponding to the historical frequency domain distortion signal, then determines the frequency response of each frequency according to the frequency domain component of each frequency in the historical frequency domain distortion signal and the frequency domain component of each frequency in the experimental frequency domain restoration signal, obtains the plurality of numerator transfer coefficients and the plurality of denominator transfer coefficients according to the frequency response of each frequency and the preset target function, constructs the frequency response transfer function, transforms the frequency response transfer function, obtains the frequency domain form function, and determines the target frequency domain restoration signal according to the amplitude attenuation coefficient and the phase offset of the frequency domain form function and the amplitude and phase of each frequency in the current frequency domain distortion signal, then inverse Fourier transforms the target frequency domain restoration signal to obtain the target restoration signal, finally performs multi-feature extraction on the target restoration signal to obtain the signal feature vector, and inputs the signal feature vector into the preset machine learning model to obtain the fault classification detection result; that is, the frequency response of each frequency is determined according to the frequency domain component of each frequency in the historical frequency domain distortion signal and the frequency domain component of each frequency in the experimental frequency domain restoration signal, the plurality of numerator transfer coefficients and the plurality of denominator transfer coefficients are obtained according to the frequency response of each frequency and the preset target function, the frequency response transfer function is constructed, and the amplitude and phase of the current frequency domain distortion signal are restored according to the amplitude attenuation coefficient and the phase offset of the frequency domain transfer function corresponding to the constructed frequency response transfer function, so that the method can effectively eliminate the nonlinear influence of the mutual inductor on the amplitude and phase of the measured signal, accurately restore the real fault information, and thus improve the detection accuracy of the hidden fault of the overhead line. BRIEF DESCRIPTION OF DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0050] Among them:

[0051] Figure 1 It is a schematic diagram of the line fault detection method based on the frequency response transfer function in the embodiments of the present application.

[0052] Figure 2 It is a schematic diagram of the line fault detection device based on the frequency response transfer function in the embodiments of the present application.

[0053] Figure 3 It is an internal structure diagram of the computer device in some embodiments. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative effort belong to the scope of protection of the present application.

[0055] With the continuous development of power systems, the safety and reliability of distribution overhead lines as a key component of the power supply network are increasingly valued. However, in actual operation, distribution overhead lines are affected by external environment and line factors, leading to frequent faults such as high-resistance grounding faults and hidden faults. Traditional fault detection methods mainly rely on direct measurement and analysis of line electrical quantities, but in a complex electromagnetic environment, the signal often appears nonlinear distortion, which makes it extremely difficult to accurately extract fault features, and further leads to low detection accuracy of hidden faults.

[0056] As a commonly used signal acquisition device in overhead lines, the non-ideal characteristics of the mutual inductor have nonlinear effects on the amplitude and phase of the measured signal, which can cause distortion of the fault signal and make the detection system unable to accurately restore the real fault information, thereby affecting the accuracy of fault detection.

[0057] To solve the above problems, the present application provides a line fault detection method based on frequency response transfer function, medium and equipment, which can effectively eliminate the nonlinear effects of the mutual inductor on the amplitude and phase of the measured signal, accurately restore the real fault information, and improve the detection accuracy of overhead line hidden faults. The specific implementation principle will be described in detail in the following embodiments.

[0058] It should be particularly noted that the present application has three stages of nonlinear effects of the non-ideal characteristics of the mutual inductor on the measured signal: the first stage is the nonlinear effect of the non-ideal characteristics of the mutual inductor on the amplitude and phase of the measured signal, which is generally caused by the characteristics of the core permeability and winding inductance of the mutual inductor; the second stage is the nonlinear effect of the non-ideal characteristics of the mutual inductor on the time domain of the measured signal, which is generally caused by the characteristics of core saturation and hysteresis; the third stage is the dynamic nonlinear effect of the non-ideal characteristics of the mutual inductor on the time domain of the measured signal, which is generally caused by the characteristics of temperature drift and core aging of the mutual inductor.

[0059] In a first aspect, the present application provides a line fault detection method based on frequency response transfer function.

[0060] Please refer to Figure 1As a schematic diagram of a line fault detection method based on a frequency response transfer function in the embodiments of the present application, the method comprises the following steps:

[0061] Step 110: obtaining a current frequency domain distortion signal and a historical frequency domain distortion signal of the mutual inductor, and an experimental frequency domain restoration signal corresponding to the historical frequency domain distortion signal.

[0062] It should be noted that the signal herein can be an electrical signal, which includes but is not limited to a current signal, a voltage signal, etc.

[0063] For the obtaining method of the current frequency domain distortion signal, the historical frequency domain distortion signal and the experimental frequency domain restoration signal, in some embodiments, the current distortion signal and the historical distortion signal of the mutual inductor can be obtained, and then the Fourier transform is performed on the current distortion signal, the historical distortion signal and the experimental restoration signal to obtain the current frequency domain distortion signal, the historical frequency domain distortion signal and the experimental frequency domain restoration signal.

[0064] It should be noted that the current distortion signal of the mutual inductor can be output by the signal output by the current and / or voltage real-time measured in the line by the mutual inductor; the historical distortion signal can be input into the mutual inductor, and then the signal output by the mutual inductor; and the experimental restoration signal can be generated by the signal generator by the operator in advance.

[0065] For the generation method of the experimental restoration signal, in some embodiments, the working frequency range of the mutual inductor can be obtained, the signal generator can be controlled to generate a signal, and the frequency of the generated signal can be adjusted in a preset step to increase sequentially in the working frequency range to obtain a signal with multiple frequencies, which is the experimental restoration signal.

[0066] For the value of the working frequency range, in some embodiments, the present application preferably sets the working frequency range to 0.1 Hz to 10 kHz.

[0067] Step 120: determining the frequency response of each frequency according to the frequency domain component of each frequency in the historical frequency domain distortion signal and the frequency domain component of each frequency in the experimental frequency domain restoration signal, and obtaining a plurality of numerator transfer coefficients and a plurality of denominator transfer coefficients according to the frequency response of each frequency and a preset target function to construct the frequency response transfer function.

[0068] The preset target function can be obtained by the operator according to a large amount of experience, experiment or statistics and pre-set, of course, it can also be pre-set by the operator according to the actual demand.

[0069] It should be noted that the preset target function is used to make the frequency response of each frequency reach the minimum, and the plurality of numerator transfer coefficients and the plurality of denominator transfer coefficients corresponding to the plurality of numerator transfer coefficients are obtained, so that the frequency response transfer function can be constructed according to the plurality of numerator transfer coefficients and the plurality of denominator transfer coefficients.

[0070] For the construction method of the preset target function, in some embodiments, the numerator and the denominator can be a plurality of polynomials of multiple orders, and the transfer function is constructed, and then the preset target function is constructed according to the transfer function and the frequency response of each frequency. The plurality of coefficients in the plurality of polynomials of multiple orders of the numerator and the plurality of coefficients in the plurality of polynomials of multiple orders of the denominator are unknown and adjustable in the transfer function, so that the plurality of coefficients in the plurality of polynomials of multiple orders of the numerator and the plurality of coefficients in the plurality of polynomials of multiple orders of the denominator are adjusted to make the frequency response of each frequency reach the minimum. The plurality of coefficients in the plurality of polynomials of multiple orders of the numerator are used as the numerator transfer coefficients, and the plurality of coefficients in the plurality of polynomials of multiple orders of the denominator are used as the denominator transfer coefficients.

[0071] For the construction method of the frequency response transfer function, in some embodiments, the plurality of numerator transfer coefficients and the plurality of denominator transfer coefficients are substituted into the transfer function, and the constructed frequency response transfer function is obtained.

[0072] Step 130: The frequency response transfer function is converted to obtain a frequency domain function, and the target frequency domain restoration signal is determined according to the amplitude attenuation coefficient and the phase offset of the frequency domain function and the amplitude and phase of each frequency in the current frequency domain distortion signal.

[0073] It should be noted that the frequency response transfer function of the present application belongs to the s domain of Laplace transform, and the frequency domain function belongs to the frequency domain of Fourier transform, so in some embodiments, the s domain and the frequency domain of the frequency response transfer function can be converted to obtain the frequency domain function. S is the Laplace operator.

[0074] It should be further noted that the non-ideal characteristics of the transformer have a nonlinear effect on the amplitude and phase of the measured signal. In this regard, the present application compensates for the frequency domain distortion of the amplitude and phase of each frequency in the current frequency domain distortion signal according to the amplitude attenuation coefficient and the phase offset of the frequency domain function, so as to obtain the target frequency domain restoration signal through the compensation of the frequency domain distortion.

[0075] Step 140: The target frequency domain restoration signal is inverse Fourier transformed to obtain a target restoration signal.

[0076] It should be noted that the current distorted signal initially measured by the transformer of the present application belongs to a time domain signal, and the target frequency domain restoration signal belongs to a frequency domain signal. Therefore, in some embodiments, inverse Fourier transform of the target frequency domain restoration signal can be performed in the frequency domain and the time domain to obtain a target restoration signal in the time domain.

[0077] Further, it should be noted that the non-ideal characteristics of the transformer not only have a nonlinear impact on the amplitude and phase of the measured signal, but also have a nonlinear impact on the time domain of the measured signal. This nonlinear impact on the time domain also causes distortion of the fault signal, making the detection system unable to accurately restore the real fault information, thereby affecting the accuracy of fault detection. Therefore, after obtaining the target restoration signal, i.e., after step 140, in other embodiments, the target restoration signal can also be compensated for time domain distortion.

[0078] That is, after obtaining the target restoration signal, in other embodiments, the historical distorted signal of the transformer and the experimental restoration signal corresponding to the historical distorted signal can be obtained, and then a polynomial model can be constructed according to the historical distorted signal and the experimental restoration signal. The data point at the t-th time point in the target restoration signal is substituted into the polynomial model to obtain the data point at the t-th time point in the first final restoration signal, where t takes an integer greater than 0 in turn until n is equal to the total number of time points of the target restoration signal, to obtain the first final restoration signal. Finally, the first final restoration signal is used as the target restoration signal again.

[0079] In some embodiments, the polynomial model can be constructed according to the data points at each time point in the historical distorted signal and the data points at each time point in the experimental restoration signal to determine a plurality of time-varying coefficients to construct the polynomial model.

[0080] In the present application, the time-varying coefficients are solved to construct the polynomial model, and then the data points at each time point in the target restoration signal are substituted into the polynomial model in turn to compensate for the time domain distortion of the data points at each time point in the target restoration signal. Through the compensation for the time domain distortion, the first final restoration signal is obtained, and finally the first final restoration signal is used as the target restoration signal again. This not only effectively eliminates the nonlinear impact of the transformer on the amplitude and phase of the measured signal, but also further eliminates the nonlinear impact of the transformer on the time domain of the measured signal, more accurately restores the real fault information, and thus improves the detection accuracy of the hidden fault of the overhead line.

[0081] Further, for the determination manner of the plurality of time-varying coefficients and the construction manner of the polynomial model, in some embodiments, a multi-order polynomial can be used, a plurality of coefficients of the multi-order polynomial can be determined according to the data points at each time in the historical distortion signal and the data points at each time in the experimental restoration signal, and the plurality of coefficients can be all used as time-varying coefficients to obtain the plurality of time-varying coefficients, and the plurality of time-varying coefficients can be substituted into the multi-order polynomial to obtain the constructed polynomial model; wherein, the order of the multi-order polynomial can be set by an operator according to a large amount of experience, experiment or statistics, and of course, the order of the multi-order polynomial can also be set by the operator according to actual needs.

[0082] Further, for the determination manner of the plurality of time-varying coefficients and the construction manner of the polynomial model, in some embodiments, a multi-order polynomial can be used, a plurality of coefficients of the multi-order polynomial can be determined according to the data points at each time in the historical distortion signal and the data points at each time in the experimental restoration signal, and the plurality of coefficients can be all used as time-varying coefficients to obtain the plurality of time-varying coefficients, and the plurality of time-varying coefficients can be substituted into the multi-order polynomial to obtain the constructed polynomial model; wherein, the order of the multi-order polynomial can be set by an operator according to a large amount of experience, experiment or statistics, and of course, the order of the multi-order polynomial can also be set by the operator according to actual needs.

[0083] In the present application, the time-varying coefficients of the polynomial model are determined by standardization processing and the least square method, which improves the construction efficiency and accuracy of the model, and further improves the effect of time-domain distortion compensation.

[0084] Further, for the determination manner of the plurality of time-varying coefficients and the construction manner of the polynomial model, in some embodiments, a multi-order polynomial can be used, a plurality of coefficients of the multi-order polynomial can be determined according to the data points at each time in the historical distortion signal and the data points at each time in the experimental restoration signal, and the plurality of coefficients can be all used as time-varying coefficients to obtain the plurality of time-varying coefficients, and the plurality of time-varying coefficients can be substituted into the multi-order polynomial to obtain the constructed polynomial model; wherein, the order of the multi-order polynomial can be set by an operator according to a large amount of experience, experiment or statistics, and of course, the order of the multi-order polynomial can also be set by the operator according to actual needs.

[0085] In the present application, the standardization processing based on the mean and the standard deviation improves the standardization and comparability of the signal data, which lays a solid foundation for the subsequent accurate determination of the time-varying coefficients and the construction of the polynomial model.

[0086] For the determination manner of the data point at the tth time in the first final restoration signal, in some embodiments, the expression of the polynomial model can be used to obtain the first final restoration signal; wherein, is the data point at the tth time in the first final restoration signal, is the n+1th time-varying coefficient, is the total number of time-varying coefficients, The data point of the tth moment in the target restored signal is obtained.

[0087] In the present application, the data points of each moment of the first final restored signal are accurately determined through the polynomial model expression, effectively improving the accuracy and reliability of signal restoration.

[0088] It should be further explained that the non-ideal characteristics of the transformer not only have nonlinear effects on the amplitude and phase of the measured signal and the time domain, but also have dynamic nonlinear effects on the time domain of the measured signal. Such dynamic nonlinear effects change with time and environment, which also causes distortion of the fault signal, making the detection system unable to accurately restore the true fault information, thereby affecting the accuracy of fault detection. Therefore, after obtaining the target restored signal, that is, after step 140, in some other embodiments, the present application can also perform dynamic compensation for the time domain distortion of the target restored signal.

[0089] That is, after obtaining the target restored signal, in some other embodiments, the historical distorted signal of the transformer and the experimental restored signal corresponding to the historical distorted signal can be obtained, and then a polynomial model is constructed according to the historical distorted signal and the experimental restored signal. The first time-varying coefficient and the second time-varying coefficient in the polynomial model are set to 0, and after taking the absolute value of the square, they are used as a residual model. The data point of the tth moment in the target restored signal is substituted into the polynomial model to obtain the data point of the tth moment in the second final restored signal. According to the residual model, the data point of the tth moment in the target restored signal, and each time-varying coefficient in the polynomial model, each time-varying coefficient in the polynomial model is updated to obtain an updated polynomial model, which is used as the polynomial model. Wherein, t takes an integer greater than 0 in turn until t is equal to the total number of moments of the target restored signal to obtain the second final restored signal. Finally, the second final restored signal is used as the target restored signal again.

[0090] It should be noted that since each time-varying coefficient in the polynomial model is updated at each moment, and the residual model is obtained based on the polynomial model, accordingly, each time-varying coefficient in the residual model is also updated with the update of each time-varying coefficient in the polynomial model.

[0091] For the construction method of the polynomial model, the construction method of the polynomial model for compensation of the time domain distortion in the above embodiments is similar, which will not be repeated here.

[0092] In the present application, by constructing a polynomial model, then setting the first time-varying coefficient and the second time-varying coefficient in the polynomial model to 0, and taking the absolute value after squaring as a residual model, substituting the data point at the tth moment in the target restoration signal into the polynomial model to obtain the data point at the tth moment in the second final restoration signal, and updating each time-varying coefficient in the polynomial model according to the residual model, the data point at the tth moment in the target restoration signal and each time-varying coefficient in the polynomial model, the updated polynomial model is used as the polynomial model again, that is, the polynomial model is dynamically updated, thereby dynamically compensating the time-domain distortion of each moment in the target restoration signal, and the second final restoration signal is obtained through the dynamic compensation of the time-domain distortion. Finally, the second final restoration signal is used as the target restoration signal again, which not only effectively eliminates the nonlinear influence of the mutual inductor on the amplitude, phase and time domain of the measured signal, but also further eliminates the dynamic nonlinear influence of the mutual inductor on the time domain of the measured signal, thereby accurately restoring the real fault information, and improving the detection accuracy of the hidden fault of the overhead line.

[0093] For the updating method of the polynomial model, in some embodiments, the data point at the tth moment in the target restoration signal can be substituted into the residual model to obtain the residual at the tth moment, and then the (n+1)th time-varying coefficient at the t+1th moment can be determined according to the residual at the tth moment, the data point at the tth moment in the target restoration signal and the nth time-varying coefficient in the polynomial model, wherein n takes an integer greater than -1 in turn until n is equal to the total number of time-varying coefficients, thereby obtaining each time-varying coefficient at the t+1th moment, and finally each time-varying coefficient in the polynomial model is updated according to each time-varying coefficient at the t+1th moment, thereby obtaining the updated polynomial model.

[0094] In the present application, by dynamically updating the time-varying coefficients of the polynomial model, time-domain dynamic nonlinear compensation of the target restoration signal is realized, the signal restoration accuracy is significantly improved, and the influence of the non-ideal characteristics of the mutual inductor on the detection accuracy is extremely eliminated.

[0095] Further, for the determination method of the residual at the tth moment, in some embodiments, the residual at the tth moment can be obtained by using the expression of the residual model ; wherein, is the residual at the tth moment, is the (n+1)th time-varying coefficient, is the total number of time-varying coefficients, is the data point at the tth moment in the target restoration signal.

[0096] In the embodiments of the present application, the residual at the tth moment is determined by using the residual model expression, which provides a key basis for dynamically updating the time-varying coefficients of the polynomial model, effectively improves the signal restoration accuracy, and further eliminates the influence of the non-ideal characteristics of the mutual inductor on the detection accuracy.

[0097] Further, for the determination manner of the nth time-varying coefficient at the t+1th moment, in some embodiments, the formula is used to determine the nth time-varying coefficient at the t+1th moment; wherein, is the nth time-varying coefficient at the t+1th moment, is the n+1th time-varying coefficient, is a preset learning rate, is the residual at the tth moment, is the data point of the target restoration signal at the tth moment.

[0098] It should be noted that the preset learning rate can be obtained by the operator according to a large amount of experience, experiment or statistics and pre-set, of course, it can also be pre-set by the operator according to the actual demand.

[0099] In the present application, the adaptive update of the time-varying coefficients of the polynomial model is realized by the dynamic formula, which significantly improves the accuracy and efficiency of the time-domain dynamic nonlinear compensation.

[0100] For the determination manner of the data point of the tth moment in the second final restoration signal, in some embodiments, the expression of the polynomial model is used to determine the second final restoration signal (the time-varying coefficients in the polynomial model of this embodiment are updated in real time); wherein, is the data point of the tth moment in the second final restoration signal, is the n+1th time-varying coefficient, is the total number of time-varying coefficients, is the data point of the tth moment in the target restoration signal.

[0101] In the present application, the data points of the second final restoration signal at each moment are accurately determined by the polynomial model expression, which significantly improves the accuracy and reliability of the time-domain dynamic nonlinear compensation, and extremely eliminates the influence of the non-ideal characteristics of the mutual inductor on the detection accuracy.

[0102] Step 150: performing multi-feature extraction on the target restoration signal to obtain a signal feature vector, and inputting the signal feature vector into a preset machine learning model to obtain a fault classification detection result.

[0103] The preset machine learning model here refers to a pre-trained model used to predict the output fault classification detection result according to the input signal feature vector.

[0104] For the signal feature vector contains the signal feature type extracted by multiple features, in some embodiments, the signal feature vector includes spectral center frequency, frequency bandwidth, short-time energy, time-domain waveform variation, pulse indicator, kurtosis, margin factor, waveform factor, frequency standard deviation, mean square frequency, energy entropy, etc.

[0105] For the acquisition method of the preset machine learning model, in some embodiments, a large number of experimental restoration signals can be obtained, multiple features are extracted from the experimental restoration signals to obtain experimental signal feature vectors, and the labels corresponding to the experimental signal feature vectors are determined, then a large number of experimental signal feature vectors and the labels corresponding thereto are input into an initial machine learning model for training to obtain the preset machine learning model; in other embodiments, a large number of historical distortion signals (which can be obtained based on experimental restoration signals, or can be obtained from line measurement) can also be obtained, and the above steps 110 to 140 are performed for signal restoration to obtain historical restoration signals, then multiple features are extracted from the historical restoration signals to obtain historical signal feature vectors, and the labels corresponding to the historical signal feature vectors are determined, finally a large number of historical signal feature vectors and the labels corresponding thereto are input into an initial machine learning model for training to obtain the preset machine learning model.

[0106] In the embodiments of the present application, the frequency response of each frequency is determined according to the frequency domain component of each frequency in the historical frequency domain distortion signal and the frequency domain component of each frequency in the experimental frequency domain restoration signal, and a plurality of numerator transfer coefficients and a plurality of denominator transfer coefficients are obtained according to the frequency response of each frequency and the preset target function to construct the frequency response transfer function, so that the amplitude and phase of the current frequency domain distortion signal are restored according to the amplitude attenuation coefficient and phase offset of the frequency domain transfer function corresponding to the constructed frequency response transfer function, which can effectively eliminate the nonlinear influence of the mutual inductor on the amplitude and phase of the measured signal, accurately restore the real fault information, and thus improve the detection accuracy of the hidden fault of the overhead line.

[0107] In addition, in addition to the above-mentioned effects of eliminating the nonlinear effects of the mutual inductor on different aspects of the signal and improving the detection accuracy of the hidden fault of the overhead line, the line fault detection method based on the frequency response transfer function has the following beneficial effects: multi-dimensional signal restoration: this method not only compensates for the frequency domain distortion of the nonlinear effects of the mutual inductor on the amplitude and phase of the signal, but also considers the static and dynamic nonlinear effects on the time domain of the signal. By constructing a polynomial model and performing corresponding compensation, the signal is restored from the frequency domain to the time domain in all directions, and the real fault information is restored to the maximum extent, providing more accurate and comprehensive data basis for fault detection; adapt to complex signal changes: the various effects caused by the non-ideal characteristics of the mutual inductor will change dynamically with time and environment. This method can adapt to these changes in real time by dynamically updating the time-varying coefficients of the polynomial model, dynamically compensating the target restored signal, and ensuring accurate signal restoration under different working conditions, thereby enhancing the processing capability of the complex and variable power system signal; optimize the construction of the polynomial model: when constructing the polynomial model for time domain distortion compensation, standardization processing and least square method are used to determine the time-varying coefficients. Standardization processing improves the standardization and comparability of signal data, and the least square method can quickly and accurately solve the model parameters, thereby improving the construction efficiency and accuracy of the model, and further improving the effect of time domain distortion compensation; adaptive model updating: in the process of time domain dynamic nonlinear compensation, the adaptive updating of the time-varying coefficients of the polynomial model is realized through a dynamic formula. The preset learning rate can be set according to the actual situation, making the model updating process more flexible and controllable. The model parameters can be dynamically adjusted according to real-time data, significantly improving the accuracy and efficiency of time domain dynamic nonlinear compensation; multi-feature extraction enhances feature richness: multi-feature extraction is performed on the target restored signal to obtain a signal feature vector containing multiple features such as center frequency, frequency bandwidth and short-time energy. Rich feature information can more comprehensively reflect the characteristics of the signal, provide more effective information for the preset machine learning model, and help improve the accuracy and reliability of the fault classification detection result; multiple training data sources ensure model generalization: the preset machine learning model can be obtained in two ways, one is to train a large number of experimentally restored signals and their labels, and the other is to train historical restored signals and their labels obtained by restoring historical distorted signals measured from the line. Multiple training data sources can enable the model to learn signal features in different scenarios, enhance the generalization ability of the model, and ensure the stability of fault detection in various complex situations; reduce dependence on professional equipment: this method mainly realizes fault detection through signal processing and analysis, which reduces the dependence on expensive equipment to a certain extent and reduces the detection cost compared with some traditional methods that require a large number of high-precision professional detection equipment;Simplify the fault detection process: the method provides a set of system complete signal restoration and fault detection process, from signal acquisition, frequency domain distortion compensation to time domain distortion compensation, and then to multi-feature extraction and fault classification detection, each link is closely connected and has clear operation steps, which is convenient for engineering and technical personnel to implement, and reduces the technical difficulty and operation complexity of fault detection.

[0108] In an implementation, the step 120 in the above embodiment comprises: determining the frequency response of each frequency according to the frequency domain component of each frequency in the historical frequency domain distortion signal and the frequency domain component of each frequency in the experimental frequency domain restoration signal, and obtaining the plurality of numerator transfer coefficients and the plurality of denominator transfer coefficients according to the frequency response of each frequency and the preset target function to construct the frequency response transfer function, including: taking the quotient value between the frequency domain component of the dth frequency in the historical frequency domain distortion signal and the frequency domain component of the dth frequency in the experimental frequency domain restoration signal as the frequency response of the dth frequency; wherein d takes an integer greater than 0 in turn until d is equal to the total number of frequencies of the historical frequency domain distortion signal or the experimental frequency domain restoration signal to obtain the frequency response of each frequency; using the LM method, adjusting the plurality of numerator transfer coefficients and the plurality of denominator transfer coefficients in the preset target function according to the frequency response of each frequency to make the function value of the preset target function reach the minimum to obtain the plurality of numerator transfer coefficients and the plurality of denominator transfer coefficients under the condition of meeting the preset target function; and constructing the frequency response transfer function according to the plurality of numerator transfer coefficients and the plurality of denominator transfer coefficients.

[0109] Wherein, LM is Levenberg-Marquardt, that is, Levenberg-Marquardt method.

[0110] In the embodiments of the present application, the construction process of the frequency response transfer function is optimized by the LM method, which significantly improves the convergence speed and accuracy of the transfer coefficient solution, and enhances the adaptability of the model to complex nonlinear distortion, providing a more reliable foundation for subsequent signal restoration and fault detection.

[0111] It can be understood that the high efficient convergence: the LM method combines the advantages of gradient descent method and Gauss-Newton method, and has a faster convergence speed when dealing with nonlinear optimization problems. Compared with the traditional iterative method, the method automatically switches to Gauss-Newton method when approaching the optimal solution by dynamically adjusting the damping factor, thereby reducing the number of iterations and shortening the calculation time, and is especially suitable for real-time processing of large-scale frequency domain data; Anti-local minimum value ability: in the case that the preset target function has multiple local minimum values, the LM method balances the global search ability of gradient descent and the local fast convergence characteristics of Gauss-Newton by introducing a damping factor, effectively reducing the risk of falling into local optimum. This feature ensures the stability of the numerator / denominator transfer coefficient solution. Even in the face of complex nonlinear distortion caused by the non-ideal characteristics of the mutual inductor, a global optimal solution can still be obtained; High-precision transfer function construction: the transfer coefficient optimized by the LM method can directly reflect the accurate characteristics of the frequency domain response of the mutual inductor. Since the method minimizes the residual (i.e. frequency response error) of the preset target function, the frequency response transfer function constructed can more accurately compensate for amplitude attenuation and phase shift, thereby significantly improving the signal restoration accuracy in the frequency domain distortion compensation stage, laying a foundation for subsequent fault feature extraction; Robustness enhancement: the LM method has low sensitivity to initial values. Even if there is a deviation in the initial transfer coefficient setting, it can still gradually approach the optimal solution through iteration. This feature makes the method more operable in practical engineering applications, and it can achieve stable convergence without relying on accurate prior knowledge, meeting the needs of different types of mutual inductors and diverse operating environments.

[0112] In a possible implementation, the expression of the preset target function in step 120 in the above embodiment is:

[0113] ;

[0114] Wherein, ;

[0115] In the above formula, is the function value of the preset target function, is the total number of frequencies of the historical frequency domain distortion signal or the experimental frequency domain restoration signal, is the frequency response of the dth frequency, is the Laplacian corresponding to the dth frequency, is the m+1th numerator transfer coefficient, is the total number of numerator transfer coefficients or denominator transfer coefficients, is the m+1th denominator transfer coefficient, is the imaginary unit, is the circular constant, is the dth frequency.

[0116] In the embodiments of the present application, by constructing a composite objective function based on frequency response, the fitting accuracy of the transfer function to the non-ideal characteristics of the transformer is significantly improved, and the anti-interference ability of the model in a complex electromagnetic environment is enhanced, thereby providing a more reliable mathematical basis for subsequent signal restoration.

[0117] It can be understood that the full-band error minimization: the objective function integrates the response errors of all frequency points in the form of summation, realizes full-band coverage from low frequency to high frequency, and such a global optimization mechanism avoids the problem of local band overfitting, ensures that the transfer function can accurately compensate for the amplitude frequency / phase frequency characteristic distortion of the transformer in the working frequency range of the transformer, and significantly improves the restoration accuracy of the characteristic frequency band such as high resistance grounding fault; complex domain dynamic weighting mechanism: the introduction of Laplace operator and imaginary unit enables the objective function to simultaneously optimize the real part (amplitude response) and the imaginary part (phase response) of the transfer function, and by adjusting the numerator / denominator coefficients, the nonlinear shift caused by the transformer can be accurately matched; anti-noise interference design: the square operation in the objective function has natural noise suppression characteristics, and in the power distribution line environment with strong electromagnetic interference, this design can effectively reduce the influence of random noise on frequency response estimation, so that the transfer function construction process has robustness to background noise, and the stability of fault feature extraction is ensured.

[0118] In a feasible implementation manner, the expression of the frequency response transfer function in step 120 in the above embodiment is:

[0119] ;

[0120] wherein, ;

[0121] In the above formula, is the frequency response transfer function, is the Laplace operator, is the m+1th numerator transfer coefficient, is the total number of numerator transfer coefficients or denominator transfer coefficients, is the m+1th denominator transfer coefficient, is the imaginary unit, is the circular constant, is the frequency.

[0122] In the embodiments of the present application, by constructing a frequency response transfer function based on the Laplace domain, high-precision mathematical modeling of the non-ideal characteristics of the transformer is realized, the amplitude / phase accuracy of signal restoration is significantly improved, and the adaptability of the model to nonlinear distortion is enhanced, thereby providing a reliable signal basis for subsequent fault feature extraction.

[0123] It can be understood that the full-band dynamic compensation capability: the transfer function adopts a multi-order polynomial structure, which can accurately fit the amplitude-frequency characteristics and phase-frequency characteristics of the mutual inductor in the working frequency range; the complex domain parameter optimization mechanism: by introducing the imaginary unit and the Laplace operator, the transfer function can optimize the real part (amplitude response) and the imaginary part (phase response) at the same time.

[0124] In a possible implementation, the expression of the frequency domain function in step 130 in the above embodiment is:

[0125] ;

[0126] wherein, is the frequency domain function, is the amplitude attenuation coefficient of the frequency f, is a natural constant, is an imaginary unit, is the phase offset of the frequency f.

[0127] In the embodiments of the present application, the independent compensation mechanism of amplitude attenuation and phase offset is realized by constructing the frequency domain function, which significantly improves the accuracy and flexibility of the frequency domain distortion compensation, and at the same time provides high-precision initial conditions for subsequent, effectively solving the distortion problem caused by the non-ideal characteristics of the mutual inductor.

[0128] It can be understood that the amplitude-phase decoupling compensation mechanism: the frequency domain function decouples the amplitude attenuation coefficient and the phase offset into independent parameters, which can accurately compensate for the amplitude-frequency characteristic distortion and the phase-frequency characteristic distortion of the mutual inductor; the complex domain dynamic modeling capability: by introducing the natural constant and the imaginary unit, the frequency domain function can be expressed in complex form, thereby representing the real part (amplitude) and the imaginary part (phase) of the signal at the same time, and this modeling method enables the model to accurately match the non-linear offset caused by the mutual inductor; the frequency band adaptive compensation characteristic: the amplitude attenuation coefficient and the phase offset are functions of the frequency, which can dynamically adapt to the differences in non-linear characteristics of the mutual inductor at different frequency bands.

[0129] In a possible implementation, the step 130 in the above embodiment determines the target frequency domain restoration signal according to the amplitude attenuation coefficient and the phase offset of the frequency domain function, and the amplitude and the phase of each frequency in the current frequency domain distorted signal, and includes:

[0130] The target frequency domain restoration signal is determined by using the formula

[0131] wherein, is the frequency domain component of the kth frequency in the target frequency domain restoration signal, is the kth frequency, is the amplitude of the kth frequency in the current frequency domain distorted signal, ​is the amplitude attenuation coefficient of the frequency f k is a natural constant, is a natural constant, is a natural constant, is the phase of the kth frequency in the current frequency domain distortion signal, is the phase offset of the frequency f k .

[0132] In the embodiments of the present application, the amplitude attenuation and phase offset caused by the transformer are accurately and independently compensated by the frequency domain component compensation formula, which significantly improves the restoration accuracy of the frequency domain distortion signal and provides a high-fidelity signal basis for subsequent fault feature extraction.

[0133] It can be understood that the amplitude-phase decoupling compensation mechanism: the formula decouples the amplitude compensation term and the phase compensation term, and can independently optimize the amplitude-frequency characteristic distortion and the phase-frequency characteristic distortion of the transformer: complex domain dynamic modeling capability: by introducing the natural constant and the imaginary unit, the formula can be expressed in complex form, thereby representing the real part (amplitude) and the imaginary part (phase) of the signal at the same time. This modeling method enables the model to accurately match the nonlinear shift caused by the transformer; frequency band adaptive compensation characteristics: the amplitude attenuation coefficient and the phase offset are functions of the frequency, which can dynamically adapt to the differences in nonlinear characteristics of the transformer in different frequency bands.

[0134] In a feasible implementation manner, the step 150 in the above embodiment performs multi-feature extraction on the target restoration signal to obtain a signal feature vector, including: performing wavelet transform on the target restoration signal to obtain a first time-frequency graph; performing short-time Fourier transform on the target restoration signal to obtain a second time-frequency graph; performing multi-feature extraction on the first time-frequency graph and the second time-frequency graph to obtain a spectral center frequency, a frequency band width, a short-time energy, and a time-domain waveform change amount; and taking the spectral center frequency, the frequency band width, the short-time energy, and the time-domain waveform change amount as the signal feature vector.

[0135] In the embodiments of the present application, the multi-time-frequency analysis method is used to fuse and extract features, which significantly improves the comprehensiveness of the fault signal feature representation and the accuracy of the fault classification detection.

[0136] It can be understood that the time-frequency joint representation enhances: combining the complementary characteristics of wavelet transform and short-time Fourier transform, both the time-frequency localized analysis capability of wavelet transform and the frequency resolution advantage of short-time Fourier transform are utilized, and a more complete signal time-frequency feature space is constructed; multi-dimensional feature fusion: four core features of spectral center frequency, frequency bandwidth, short-time energy and time-domain waveform change are extracted, forming a multi-dimensional feature vector covering the frequency domain, time domain and energy domain, effectively solving the problem of insufficient representation of single feature to complex fault mode; anti-interference ability is improved: time-frequency analysis method has natural adaptability to non-stationary fault signal, and noise interference can be suppressed through time-frequency graph processing, and the extracted frequency bandwidth and time-domain waveform change have higher sensitivity to hidden faults; the feature interpretability is enhanced: the four extracted features have clear physical meaning and strong correlation with fault types, providing interpretable input features for machine learning models, which helps to improve the credibility of fault classification results; the calculation efficiency is optimized: the standardized time-frequency transformation process is adopted, and both wavelet transform and short-time Fourier transform can be realized through fast algorithm, and the feature extraction process maintains linear complexity, which meets the real-time detection demand while ensuring the feature quality.

[0137] In a possible implementation, the step 110 in the above embodiment of acquiring the current frequency domain distortion signal of the mutual inductor comprises: acquiring a current distortion signal of the mutual inductor; performing Fourier transform on the current distortion signal to obtain an initial frequency domain distortion signal; determining a maximum frequency according to the experimental frequency domain restoration signal; determining a cutoff frequency according to the maximum frequency and a preset safety coefficient; and performing frequency screening on the initial frequency domain distortion signal according to the cutoff frequency to obtain the current frequency domain distortion signal.

[0138] The preset safety coefficient can be set in advance by an operator according to a large amount of experience, experiments or statistics, or can be set in advance by the operator according to actual needs.

[0139] For the value of the preset safety coefficient, in some embodiments, the preset safety coefficient is preferably set to 1.2.

[0140] For the determination method of the maximum frequency, in some embodiments, the maximum frequency can be determined as the maximum frequency among the frequency domain components of each frequency of the experimental frequency domain restoration signal; in other embodiments, the working frequency range of the mutual inductor can be used to determine the maximum frequency, for example, when the working frequency range of the mutual inductor is 0.1 Hz to 10 kHz, 10 kHz is used as the maximum frequency.

[0141] For the determination method of the cutoff frequency, in some embodiments, the product of the maximum frequency and the preset safety coefficient can be used as the cutoff frequency.

[0142] For the determination manner of the current frequency domain distortion signal, in some embodiments, among the frequency domain components of each frequency of the initial frequency domain distortion signal, the frequency domain components corresponding to the frequencies greater than the cutoff frequency are removed to obtain the current frequency domain distortion signal.

[0143] In the embodiments of the present application, high-frequency noise interference is effectively suppressed through the frequency screening mechanism, and the signal-to-noise ratio of the fault signal and the accuracy of subsequent feature extraction are significantly improved.

[0144] It can be understood that the adaptive frequency band constraint: according to the experimental frequency domain restoration signal, the maximum frequency is determined, and the cutoff frequency is set in combination with a safety factor of 1.2, which not only retains the effective fault feature frequency band, but also avoids the pollution of high-frequency noise to the signal, and is more adaptable than the fixed frequency band scheme; the anti-aliasing effect is improved: after the Fourier transform, the components outside the cutoff frequency are accurately removed, which effectively prevents the aliasing of high-frequency noise into the baseband frequency band, and provides a purer input signal for subsequent frequency domain distortion compensation; the mutual inductor characteristic adaptation: according to the difference of the working frequency band of different mutual inductors, the cutoff frequency mechanism can automatically match the device characteristics, avoiding the problems of feature loss or noise introduction caused by the fixed frequency band.

[0145] The present application provides a line fault detection device based on frequency response transfer function in the second aspect.

[0146] Please refer to Figure 2 , the schematic diagram of the line fault detection device based on frequency response transfer function in the embodiments of the present application, the device 210 comprises:

[0147] The acquisition module 211 is configured to acquire the current frequency domain distortion signal and the historical frequency domain distortion signal of the mutual inductor, and the experimental frequency domain restoration signal corresponding to the historical frequency domain distortion signal.

[0148] The determination and construction module 212 is configured to determine the frequency response of each frequency according to the frequency domain component of each frequency in the historical frequency domain distortion signal and the frequency domain component of each frequency in the experimental frequency domain restoration signal, and obtain a plurality of numerator transfer coefficients and a plurality of denominator transfer coefficients according to the frequency response of each frequency and a preset target function, so as to construct the frequency response transfer function.

[0149] The conversion and determination module 213 is configured to convert the frequency response transfer function to obtain a frequency domain form function, and determine the target frequency domain restoration signal according to the amplitude attenuation coefficient and the phase offset of the frequency domain form function, and the amplitude and phase of each frequency in the current frequency domain distortion signal.

[0150] The inverse transform module 214 is configured to perform inverse Fourier transform on the target frequency domain restoration signal to obtain a target restoration signal.

[0151] The extraction and prediction module 215 is configured to perform multi-feature extraction on the target restored signal to obtain a signal feature vector, input the signal feature vector into a preset machine learning model, and obtain a fault classification detection result.

[0152] In the embodiments of the present application, the related content of the above-mentioned acquisition module 211, the determination and construction module 212, the conversion and determination module 213, the inverse transformation module 214 and the extraction and prediction module 215 can be referred to the content in the above-mentioned embodiments, and details are not described herein. Figure 1

[0153] It should be noted that the device 210 of the present application further includes other modules, and it can be understood that the method of the present application has a one-to-one correspondence with the device 210, so the other modules of the device 210 of the present application are the corresponding content of the method of the present application in the above-mentioned embodiments.

[0154] In the embodiments of the present application, by determining the frequency response of each frequency according to the frequency domain component of each frequency in the historical frequency domain distortion signal and the frequency domain component of each frequency in the experimental frequency domain restored signal, and obtaining a plurality of numerator transfer coefficients and a plurality of denominator transfer coefficients according to the frequency response of each frequency and a preset target function, a frequency response transfer function is constructed, so that the amplitude and phase of the current frequency domain distortion signal are restored according to the amplitude attenuation coefficient and the phase shift of the frequency domain transfer function corresponding to the constructed frequency response transfer function. The device can effectively eliminate the nonlinear influence of the mutual inductor on the amplitude and phase of the measured signal, accurately restore the real fault information, and thus improve the detection accuracy of the hidden fault of the overhead line.

[0155] ​In addition, in addition to the above-mentioned effects of eliminating the nonlinear effects of the mutual inductor on different aspects of the signal and improving the detection accuracy of the hidden fault of the overhead line, the line fault detection device based on the frequency response transfer function has the following beneficial effects: multi-dimensional signal restoration: the device not only compensates for the frequency domain distortion of the nonlinear effects of the mutual inductor on the amplitude and phase of the signal, but also considers the static and dynamic nonlinear effects on the time domain of the signal. By constructing a polynomial model and performing corresponding compensation, the device realizes all-round signal restoration from the frequency domain to the time domain, maximally restores the real fault information, and provides more accurate and comprehensive data basis for fault detection; adapt to complex signal changes: the various effects caused by the non-ideal characteristics of the mutual inductor will dynamically change with time and environment. The device can adapt to these changes in real time by dynamically updating the time-varying coefficients of the polynomial model, dynamically compensating the target restored signal, and ensuring accurate signal restoration under different working conditions, thereby enhancing the processing capability of the complex and variable power system signal; optimize the construction of the polynomial model: when constructing the polynomial model for time domain distortion compensation, standardized processing and least square method are used to determine the time-varying coefficients. The standardized processing improves the standardization and comparability of the signal data, and the least square method can quickly and accurately solve the model parameters, thereby improving the construction efficiency and accuracy of the model and further improving the effect of time domain distortion compensation; adaptive model updating: in the process of time domain dynamic nonlinear compensation, the adaptive updating of the time-varying coefficients of the polynomial model is realized through a dynamic formula. The preset learning rate can be set according to the actual situation, so that the model updating process is more flexible and controllable. The model parameters can be dynamically adjusted according to the real-time data, which significantly improves the accuracy and efficiency of the time domain dynamic nonlinear compensation; multi-feature extraction enhances feature richness: multi-feature extraction is performed on the target restored signal to obtain a signal feature vector containing multiple features such as center frequency, frequency bandwidth and short-time energy. The rich feature information can more comprehensively reflect the characteristics of the signal, provide more effective information for the preset machine learning model, and help improve the accuracy and reliability of the fault classification and detection result; multiple training data sources ensure model generalization: the preset machine learning model can be obtained in two ways, one is to train a large number of experimentally restored signals and their labels, and the other is to train the historical restored signals and their labels obtained by restoring the historical distorted signals measured from the line. Multiple training data sources can enable the model to learn the signal features in different scenarios, enhance the generalization ability of the model, and ensure the stability of fault detection in various complex situations; reduce the dependence on professional equipment: the device mainly realizes fault detection through signal processing and analysis, which reduces the dependence on expensive equipment to a certain extent and reduces the detection cost compared with some traditional devices that require a large number of high-precision professional detection equipment;Simplify fault detection process: the device provides a set of system complete signal restoration and fault detection process, from signal acquisition, frequency domain distortion compensation to time domain distortion compensation, and then to multi-feature extraction and fault classification detection, each link is closely connected and has clear operation steps, which is convenient for engineering and technical personnel to implement, and reduces the technical difficulty and operation complexity of fault detection.

[0156] The application provides a computer readable storage medium in a third aspect, which stores a computer program. The computer program is executed by a processor to enable the processor to perform the line fault detection method based on the frequency response transfer function according to any one of the first aspect.

[0157] The application provides a computer device in a fourth aspect, which includes a memory and a processor. The memory stores a computer program. The computer program is executed by the processor to enable the processor to perform the line fault detection method based on the frequency response transfer function according to any one of the first aspect.

[0158] Figure 3 The internal structure diagram of the computer device in some embodiments is shown. The computer device can be a terminal, a server, or a gateway. As shown in the figure, Figure 3 The computer device includes a processor, a memory and a network interface connected through a system bus.

[0159] The memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium of the computer device stores an operating system, and can also store a computer program. The computer program is executed by the processor to enable the processor to implement each step in the above method embodiments. The internal memory can also store a computer program. The computer program is executed by the processor to enable the processor to perform each step in the above method embodiments. Those skilled in the art can understand, Figure 3 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the application, and does not constitute a limitation on the computer device to which the scheme of the application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0160] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing related hardware. The program can be stored in a non-volatile computer readable storage medium. The program, when executed, can include the processes of the above-mentioned embodiment methods.

[0161] Any reference to storage, memory, database or other medium herein can include non-volatile and / or volatile storage. Non-volatile storage can include read-only memory (ROM), programmable ROM (PROM), electronically programmable ROM (EPROM), or electrically erasable ROM (EEPROM). Volatile storage can include random-access memory (RAM). By way of illustration, and not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct Rambus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM). The RAM can also include a basic-0 type (B0-RAM), shadow RAM, and / or electromagnetic RAM (EMRAM).

[0162] Any of the technical features of the above embodiments can be combined, and for brevity, not all possible combinations of the technical features of the above embodiments are described, however, any combination of the technical features should be considered as within the scope of the present disclosure as long as the combination does not result in a contradiction.

[0163] The above embodiments merely express several implementation manners of the present application, and the description is relatively specific and detailed, but should not be understood as a limitation on the patent scope of the present application. It should be noted that, for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present application, and these should be considered as within the protection scope of the present application. Therefore, the patent protection scope of the present application should be subject to the appended claims.

Claims

1. A line fault detection method based on frequency response transfer function, characterized in that, The method includes: Obtain the current frequency domain distortion signal and historical frequency domain distortion signal of the mutual inductor, as well as the experimental frequency domain reconstruction signal corresponding to the historical frequency domain distortion signal; Based on the frequency domain components of each frequency in the historical frequency domain distortion signal and the frequency domain components of each frequency in the experimental frequency domain reconstruction signal, the frequency response of each frequency is determined, and based on the frequency response of each frequency and the preset objective function, multiple numerator transfer coefficients and multiple denominator transfer coefficients are obtained to construct the frequency response transfer function. The frequency response transfer function is transformed to obtain a frequency domain form function, and the target frequency domain restored signal is determined based on the amplitude attenuation coefficient and phase offset of the frequency domain form function, as well as the amplitude and phase of each frequency in the current frequency domain distorted signal. Perform an inverse Fourier transform on the target frequency domain reconstructed signal to obtain the target reconstructed signal; The target restored signal is subjected to multi-feature extraction to obtain a signal feature vector, and the signal feature vector is input into a preset machine learning model to obtain a fault classification and detection result; in, The process involves determining the frequency response of each frequency based on the frequency domain components of each frequency in the historical frequency domain distortion signal and the frequency domain components of each frequency in the experimental frequency domain restored signal, and obtaining multiple numerator transfer coefficients and multiple denominator transfer coefficients based on the frequency responses of each frequency and a preset objective function to construct a frequency response transfer function, including: The quotient between the frequency component of the d-th frequency in the historical frequency-domain distorted signal and the frequency component of the d-th frequency in the experimental frequency-domain restored signal is taken as the frequency response of the d-th frequency; wherein, d takes integer values ​​greater than 0 in turn until d is equal to the total number of frequencies in the historical frequency-domain distorted signal or the experimental frequency-domain restored signal, so as to obtain the frequency response of each frequency. Using the LM method, under the condition of satisfying the preset objective function, the numerator transfer coefficients and denominator transfer coefficients in the preset objective function are adjusted according to the frequency response of each frequency so that the function value of the preset objective function is minimized, resulting in multiple numerator transfer coefficients and multiple denominator transfer coefficients. The frequency response transfer function is constructed based on the transfer coefficients of each numerator and each transfer coefficient of each denominator. The expression for the preset objective function is: ； in, ; In the above formula, The function value of the preset objective function. The total number of frequencies in the historical frequency domain distorted signal or the experimental frequency domain restored signal. Let d be the frequency response at the d-th frequency. Let d be the Laplace operator corresponding to the d-th frequency. The transfer coefficient of the (m+1)th molecule. Subtract 1 from the total number of numerator or denominator transfer coefficients. The transitivity coefficient of the (m+1)th denominator. The imaginary unit, Pi For the d-th frequency; The step of determining the target frequency domain reconstructed signal based on the amplitude attenuation coefficient and phase offset of the frequency domain form function, and the amplitude and phase of each frequency in the current frequency domain distorted signal, includes: Using formula Determine the target frequency domain reconstructed signal; in, The frequency domain component of the k-th frequency in the target frequency domain reconstructed signal. For the kth frequency, Let x be the amplitude of the k-th frequency in the current frequency domain distorted signal. For frequency f k The amplitude attenuation coefficient, It is a natural constant. The imaginary unit, The phase of the k-th frequency in the current frequency domain distorted signal. For frequency f k The phase offset.

2. The line fault detection method based on frequency response transfer function according to claim 1, characterized in that, The expression for the frequency response transfer function is: ； in, ; In the above formula, The frequency response transfer function is... For the Laplace operator, The transfer coefficient of the (m+1)th molecule. Subtract 1 from the total number of numerator or denominator transfer coefficients. The transitivity coefficient of the (m+1)th denominator. The imaginary unit, Pi For frequency.

3. The line fault detection method based on frequency response transfer function according to claim 1, characterized in that, The expression for the frequency domain form function is: ； in, For the frequency domain form function, The amplitude attenuation coefficient at frequency f is... It is a natural constant. The imaginary unit, The phase offset is the frequency f.

4. The line fault detection method based on frequency response transfer function according to claim 1, characterized in that, The step of extracting multiple features from the target restored signal to obtain a signal feature vector includes: Perform wavelet transform on the target restored signal to obtain the first time-frequency diagram; Perform a short-time Fourier transform on the target restored signal to obtain a second time-frequency diagram; Multi-feature extraction is performed on the first time-frequency graph and the second time-frequency graph to obtain the spectrum center frequency, bandwidth, short-time energy and time-domain waveform change. The center frequency of the spectrum, the bandwidth, the short-time energy, and the time-domain waveform change are used as the signal feature vector.

5. The line fault detection method based on frequency response transfer function according to claim 1, characterized in that, The acquisition of the current frequency domain distortion signal of the mutual inductor includes: Obtain the current distortion signal of the current transformer; Perform a Fourier transform on the current distorted signal to obtain the initial frequency domain distorted signal; The maximum frequency is determined based on the frequency domain reconstruction signal from the experiment. The cutoff frequency is determined based on the maximum frequency and the preset safety factor; Based on the cutoff frequency, the initial frequency domain distortion signal is frequency-filtered to obtain the current frequency domain distortion signal.

6. A computer-readable storage medium, characterized in that, The system contains a computer program that, when executed by a processor, causes the processor to perform the line fault detection method based on the frequency response transfer function as described in any one of claims 1 to 5.

7. A computer device, characterized in that, The system includes a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the line fault detection method based on the frequency response transfer function as described in any one of claims 1 to 5.

Images