Partial discharge signal extraction method and system based on frequency domain intelligent filtering

Abstract

The invention provides a partial discharge signal extraction method and system based on frequency domain intelligent filtering. The method comprises the following steps: acquiring specified data of continuous partial discharge signals of high-voltage electrical equipment; establishing a normalized spectrum distribution model of the partial discharge signal by using specified data; based on the normalized spectrum distribution model, determining a spectrum distribution interval and a first spectrum characteristic parameter of the partial discharge signal; filtering the partial discharge signal by using the spectrum distribution interval and the specified data; determining a second spectrum characteristic parameter of the filtered partial discharge signal; and based on the first frequency spectrum characteristic parameter and the second frequency spectrum characteristic parameter, rejecting the partial discharge signal to extract and obtain the partial discharge signal after interference suppression. According to the scheme, the frequency domain characteristic comparison analysis is performed by using the frequency spectrum characteristic parameters, so that the interference signal in the partial discharge signal detection process is suppressed, the partial discharge signal subjected to interference suppression is finally extracted and obtained, and the accuracy and reliability of partial discharge signal detection are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of partial discharge detection of high-voltage electrical equipment, and in particular to a method and system for extracting partial discharge signals based on frequency domain intelligent filtering. Background Art

[0002] When high-voltage electrical equipment produces partial discharge, changes in the local electromagnetic field inside the equipment will excite high-frequency electromagnetic waves (i.e., ultra-high frequency partial discharge electromagnetic signals, or partial discharge signals for short). Statistical characteristic analysis of partial discharge signals combined with voltage power frequency phase can identify defect categories.

[0003] However, in actual measurement environments, the randomness of background noise and the diversity of various interferences will affect the reliable detection and identification of partial discharge signals. Therefore, there is an urgent need for a method that can accurately detect partial discharge signals. Summary of the invention

[0004] In view of this, an embodiment of the present invention provides a method and system for extracting partial discharge signals based on frequency domain intelligent filtering, so as to accurately detect partial discharge signals.

[0005] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0006] A first aspect of an embodiment of the present invention discloses a method for extracting a partial discharge signal based on frequency domain intelligent filtering, the method comprising:

[0007] Acquire specified data of a continuous partial discharge signal of a high-voltage electrical device, wherein the specified data includes at least: pulse width, sampling interval, center frequency, number of signals and amplitude;

[0008] Using the specified data, establishing a normalized spectrum distribution model of the partial discharge signal;

[0009] Based on the normalized spectrum distribution model, determining the spectrum distribution interval and the first spectrum characteristic parameter of the partial discharge signal;

[0010] Filtering the partial discharge signal using the frequency spectrum distribution interval and the designated data;

[0011] Determining a second frequency spectrum characteristic parameter of the filtered partial discharge signal;

[0012] Based on the first frequency spectrum characteristic parameter and the second frequency spectrum characteristic parameter, the partial discharge signal is eliminated to extract the partial discharge signal with interference suppression completed.

[0013] Preferably, using the specified data to establish a normalized spectrum distribution model of the partial discharge signal includes:

[0014] Establishing a time domain model of the partial discharge signal by using the pulse width, the sampling interval and the center frequency;

[0015] Based on the time domain model of the partial discharge signal, the signal number and Gaussian white noise, combined with fast Fourier transform, determining the frequency spectrum characteristics of the partial discharge signal;

[0016] Based on the frequency spectrum characteristics of the partial discharge signal and in combination with the amplitude peak value of the frequency spectrum characteristics of the partial discharge signal, a normalized frequency spectrum distribution model of the partial discharge signal is established.

[0017] Preferably, determining the spectrum distribution interval and the first spectrum characteristic parameter of the partial discharge signal based on the normalized spectrum distribution model includes:

[0018] Using the normalized spectrum distribution model and threshold parameters, estimating the starting frequency and the cutoff frequency of the partial discharge signal;

[0019] Determining a frequency spectrum distribution interval of the partial discharge signal based on a frequency disturbance coefficient, the starting frequency and the cut-off frequency;

[0020] Based on the normalized spectrum distribution model, the start frequency and the cutoff frequency, a first spectrum characteristic parameter of the partial discharge signal is determined.

[0021] Preferably, the first spectrum characteristic parameter includes at least skewness, steepness and pulse waveform root mean square error;

[0022] Determining a first spectrum characteristic parameter of the partial discharge signal based on the normalized spectrum distribution model, the start frequency, and the cutoff frequency includes:

[0023] Determining a bandwidth and a mean frequency component of the spectrum of the partial discharge signal based on the normalized spectrum distribution model, the start frequency and the cutoff frequency;

[0024] The frequency bandwidth and the mean frequency component are used to determine the root mean square error, skewness and steepness of the pulse waveform of the partial discharge signal.

[0025] Preferably, filtering the partial discharge signal using the spectrum distribution interval and the designated data comprises:

[0026] Using the specified data, respectively establish a same-frequency interference signal duration model and a variable-frequency interference signal duration model;

[0027] Establishing a mixed signal model based on the co-frequency interference signal duration model, the variable frequency interference signal duration model and the continuous partial discharge signal;

[0028] Performing a fast Fourier transform on the mixed signal model to obtain a frequency spectrum feature of the mixed signal model;

[0029] The frequency spectrum characteristics of the mixed signal model are filtered using a bandpass filter set by the frequency spectrum distribution interval to obtain the filtered partial discharge signal.

[0030] Preferably, based on the first spectrum characteristic parameter and the second spectrum characteristic parameter, the partial discharge signal is eliminated to extract the partial discharge signal for completing interference suppression, comprising:

[0031] Calculating a parameter difference between the first frequency spectrum characteristic parameter and the second frequency spectrum characteristic parameter;

[0032] For each of the partial discharge signals, if the parameter difference corresponding to the partial discharge signal satisfies a rejection condition, the partial discharge signal is rejected;

[0033] If the parameter difference corresponding to the partial discharge signal does not satisfy the elimination condition, it is determined that the partial discharge signal is the partial discharge signal that completes interference suppression.

[0034] A second aspect of an embodiment of the present invention discloses a partial discharge signal extraction system based on frequency domain intelligent filtering, the system comprising:

[0035] An acquisition unit, used to acquire specified data of a continuous partial discharge signal of a high-voltage electrical device, wherein the specified data at least includes: pulse width, sampling interval, center frequency, number of signals and amplitude;

[0036] An establishing unit, used to establish a normalized spectrum distribution model of the partial discharge signal using the specified data;

[0037] A first determining unit, configured to determine a spectrum distribution interval and a first spectrum characteristic parameter of the partial discharge signal based on the normalized spectrum distribution model;

[0038] A filtering unit, configured to filter the partial discharge signal using the frequency spectrum distribution interval and the designated data;

[0039] A second determining unit, used to determine a second frequency spectrum characteristic parameter of the filtered partial discharge signal;

[0040] The elimination unit is used to eliminate the partial discharge signal based on the first spectrum characteristic parameter and the second spectrum characteristic parameter to extract the partial discharge signal with interference suppression completed.

[0041] Preferably, the establishing unit comprises:

[0042] A first establishing module, used to establish a time domain model of the partial discharge signal by using the pulse width, the sampling interval and the center frequency;

[0043] A determination module, used to determine the frequency spectrum characteristics of the partial discharge signal based on the time domain model of the partial discharge signal, the signal number and Gaussian white noise in combination with fast Fourier transform;

[0044] The second establishing module is used to establish a normalized spectrum distribution model of the partial discharge signal based on the spectrum characteristics of the partial discharge signal and in combination with the amplitude peak value of the spectrum characteristics of the partial discharge signal.

[0045] Preferably, the first determining unit includes:

[0046] An estimation module, used to estimate the starting frequency and the cutoff frequency of the partial discharge signal by using the normalized spectrum distribution model and the threshold parameter;

[0047] A first determination module, configured to determine a frequency spectrum distribution interval of the partial discharge signal based on a frequency disturbance coefficient, the start frequency and the cutoff frequency;

[0048] The second determination module is used to determine a first spectrum characteristic parameter of the partial discharge signal based on the normalized spectrum distribution model, the start frequency and the cutoff frequency.

[0049] Preferably, the first spectrum characteristic parameter includes at least skewness, steepness and pulse waveform root mean square error; the second determination module is specifically used for:

[0050] Determining a bandwidth and a mean frequency component of the spectrum of the partial discharge signal based on the normalized spectrum distribution model, the start frequency and the cutoff frequency;

[0051] The frequency bandwidth and the mean frequency component are used to determine the root mean square error, skewness and steepness of the pulse waveform of the partial discharge signal.

[0052] Based on the above-mentioned embodiment of the present invention, a method and system for extracting partial discharge signals based on frequency domain intelligent filtering is provided. The method is as follows: obtaining designated data of continuous partial discharge signals of high-voltage electrical equipment; using the designated data, establishing a normalized spectrum distribution model of the partial discharge signal; based on the normalized spectrum distribution model, determining the spectrum distribution interval and the first spectrum characteristic parameter of the partial discharge signal; using the spectrum distribution interval and the designated data to filter the partial discharge signal; determining the second spectrum characteristic parameter of the filtered partial discharge signal; based on the first spectrum characteristic parameter and the second spectrum characteristic parameter, eliminating the partial discharge signal to extract the partial discharge signal with interference suppression. This scheme uses the spectrum characteristic parameters to perform frequency domain feature comparison and analysis, thereby suppressing the interference signal in the partial discharge signal detection process, and finally extracting the partial discharge signal with interference suppression, thereby improving the accuracy and reliability of detecting partial discharge signals. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0054] Figure 1 A flowchart of a method for extracting partial discharge signals based on frequency domain intelligent filtering provided by an embodiment of the present invention;

[0055] Figure 2 A flow chart of establishing a normalized spectrum distribution model provided by an embodiment of the present invention;

[0056] Figure 3 A flow chart for determining a spectrum distribution interval and a first spectrum characteristic parameter provided by an embodiment of the present invention;

[0057] Figure 4 A flow chart of filtering a partial discharge signal provided by an embodiment of the present invention;

[0058] Figure 5 An example diagram of the frequency spectrum characteristics of a continuous partial discharge signal provided by an embodiment of the present invention;

[0059] Figure 6 An example diagram of the frequency spectrum characteristics of an interference signal provided by an embodiment of the present invention;

[0060] Figure 7 An example diagram of time-domain aligned three-dimensional distribution of continuous partial discharge signals and interference signals provided in an embodiment of the present invention;

[0061] Figure 8An example diagram of three-dimensional distribution of a continuous partial discharge signal after interference suppression provided by an embodiment of the present invention;

[0062] Fig. 9 A structural block diagram of a partial discharge signal extraction system based on frequency domain intelligent filtering provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0063] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0064] In this application, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

[0065] The stability and integration of gas-insulated high-voltage electrical equipment (such as GIS and GIL) make it widely used in current power development and construction, supporting long-distance cross-regional power transmission and making up for the uneven distribution and low utilization of power resources.

[0066] Under the actual operating conditions of high-voltage electrical equipment, problems such as aging of insulation components caused by long-term operation, equipment process problems, and metal foreign matter generated by movable parts will destroy the uniformity of the electric field inside the equipment and cause local discharge (PD). Local discharge will reduce the insulation margin of the equipment and affect the operating reliability of the high-voltage electrical equipment.

[0067] When high-voltage electrical equipment produces partial discharge, changes in the local electromagnetic field inside the equipment will excite high-frequency electromagnetic waves (i.e., ultra-high frequency partial discharge electromagnetic signals, referred to as ultra-high frequency partial discharge signals or partial discharge signals). Statistical characteristics analysis of partial discharge signals combined with voltage power frequency phase can be used to identify defect categories. However, in actual measurement environments, the randomness of background noise and the diversity of various types of interference will affect the reliable detection and identification of partial discharge signals.

[0068] Research has found that the current interference suppression method for partial discharge signal acquisition usually uses background channel cancellation technology. In basic research, wavelet transforms with different kernel functions and envelope feature parameter clustering analysis are widely used. These related studies all perform complex signal processing or multi-channel time series comparison analysis on the original signal. The limitations of the application and the complexity of the algorithm make interference suppression more difficult and hardware resources difficult to effectively guarantee. When identifying and judging partial discharge signals, parameters such as spectral characteristics and discharge frequency are key parameters for measuring the partial discharge state of the equipment. The generation of interference signals seriously affects the relevant parameters of the actual partial discharge signal, making it difficult to effectively perceive and evaluate the actual operating state of the equipment.

[0069] Specifically, when partial discharge occurs inside high-voltage electrical equipment, in order to further analyze the partial discharge signal, it is necessary to identify the spectrum of the partial discharge signal and locate the signal source. However, under actual working conditions, there is serious electromagnetic interference in the acquisition environment, which will lead to abnormal partial discharge signal processing and make it difficult to evaluate the stable operating status of the high-voltage electrical equipment.

[0070] Therefore, in order to effectively suppress environmental interference during partial discharge detection, this scheme proposes a partial discharge signal extraction method and system based on frequency domain intelligent filtering, and uses frequency spectrum feature parameters to perform frequency domain feature comparison and analysis, so as to suppress the interference signal in the partial discharge signal detection process, and finally extract the partial discharge signal with interference suppression, thereby improving the accuracy and reliability of the detected partial discharge signal, and thus providing strong support for the operation status evaluation of high-voltage electrical equipment.

[0071] See also Figure 1 , shows a flow chart of a partial discharge signal extraction method based on frequency domain intelligent filtering provided by an embodiment of the present invention, the partial discharge signal extraction method comprising:

[0072] Step S101: Acquire designated data of continuous partial discharge signals of high-voltage electrical equipment.

[0073] In the specific implementation of step S101, the specified data of the continuous partial discharge signal (ie, the continuous discharge signal) of the high-voltage electrical equipment is obtained, and the specified data at least includes: the pulse width (duration, denoted as T p ), sampling interval (T s )、center frequency(f c ), number of signals (M) and amplitude (amplitude parameter, denoted as A).

[0074] Step S102: using the specified data, establishing a normalized spectrum distribution model of the partial discharge signal.

[0075] In the process of specifically implementing step S102, a normalized spectral distribution model S(f) of continuous partial discharge signals is established by using specified data of continuous partial discharge signals. Among them, regarding how to establish the normalized spectral distribution model S(f), it will be described in detail in the subsequent embodiments of the present invention Figure 2 will be described in detail.

[0076] Step S103: Based on the normalized spectral distribution model, determine the spectral distribution interval and the first spectral characteristic parameters of the partial discharge signal.

[0077] In the process of specifically implementing step S103, based on the normalized spectral distribution model, determine the spectral distribution interval and the first spectral characteristic parameters of the partial discharge signal. The first spectral characteristic parameters at least include skewness, kurtosis, and root mean square error of the pulse waveform.

[0078] Among them, the skewness in the first spectral characteristic parameters is denoted as p s , the kurtosis in the first spectral characteristic parameters is denoted as z s , and the root mean square error of the pulse waveform in the first spectral characteristic parameters is denoted as .

[0079] Regarding how to determine the spectral distribution interval and the first spectral characteristic parameters, it will be described in detail in the subsequent embodiments of the present invention Figure 3 will be described in detail.

[0080] Step S104: Filter the partial discharge signal by using the spectral distribution interval and the specified data.

[0081] In the process of specifically implementing step S104, filter the partial discharge signal by using the spectral distribution interval and the specified data to obtain the filtered partial discharge signal. Among them, regarding how to filter the partial discharge signal, it will be described in detail in the subsequent embodiments of the present invention Figure 4 will be described in detail.

[0082] Step S105: Determine the second spectral characteristic parameters of the filtered partial discharge signal.

[0083] In the process of specifically implementing step S105, after obtaining the filtered partial discharge signal, determine the second spectral characteristic parameters of the filtered partial discharge signal. The second spectral characteristic parameters at least include skewness, kurtosis, and root mean square error of the pulse waveform.

[0084] Among them, the skewness in the second spectral characteristic parameters is denoted as p all , the kurtosis in the second spectral characteristic parameters is denoted as z all , and the root mean square error of the pulse waveform in the second spectral characteristic parameters is denoted as .

[0085] Step S106: based on the first spectrum characteristic parameter and the second spectrum characteristic parameter, the partial discharge signal is eliminated to extract the partial discharge signal with interference suppression completed.

[0086] In the specific implementation of step S106, the first spectrum characteristic parameter and the second spectrum characteristic parameter are used to eliminate the partial discharge signal, so as to extract the interference suppression result of the partial discharge signal, and the interference suppression result includes the partial discharge signal with completed interference suppression.

[0087] In some embodiments, a specific method for obtaining a partial discharge signal that completes interference suppression is: calculating a parameter difference between a first frequency spectrum characteristic parameter and a second frequency spectrum characteristic parameter.

[0088] Specifically, the parameter differences between the skewness, steepness and pulse waveform root mean square error in the first spectrum characteristic parameters and the skewness, steepness and pulse waveform root mean square error in the second spectrum characteristic parameters are calculated respectively.

[0089] The parameter difference between the skewness in the first spectrum characteristic parameter and the skewness in the second spectrum characteristic parameter is recorded as Δp, Δp=p all -p s ;

[0090] The parameter difference between the pulse waveform root mean square error in the first spectrum characteristic parameter and the pulse waveform root mean square error in the second spectrum characteristic parameter is recorded as ΔS 2 , ΔS 2 = - ;

[0091] The parameter difference between the steepness in the first spectrum characteristic parameter and the steepness in the second spectrum characteristic parameter is recorded as Δz, Δz=z all -z s .

[0092] For each partial discharge signal, if the parameter difference (Δp, ΔS 2 and Δz) meet the rejection conditions, then the partial discharge signal is rejected;

[0093] If the parameter difference corresponding to the partial discharge signal (Δp, ΔS 2 and Δz) do not meet the rejection condition, the partial discharge signal is determined to be a partial discharge signal that completes interference suppression.

[0094] Among them, the elimination condition is {Δp, ΔS 2 , Δz}>x, where x is the error ratio.

[0095] In an embodiment of the present invention, frequency domain feature comparison and analysis is performed using spectrum feature parameters, thereby suppressing interference signals in the process of partial discharge signal detection, and finally extracting partial discharge signals with interference suppression, thereby improving the accuracy and reliability of detecting partial discharge signals.

[0096] For the above-mentioned embodiment of the present invention Figure 1 The normalized spectrum distribution model involved in step S102 is shown in Figure 2 , shows a flow chart of establishing a normalized spectrum distribution model provided by an embodiment of the present invention, Figure 2 The steps include:

[0097] Step S201: Establish a time domain model of a partial discharge signal using pulse width, sampling interval and center frequency.

[0098] In the specific implementation of step S201, the pulse width T p and sampling interval T s , calculate the time series of the partial discharge signal (denoted as t Tp ); using amplitude A, time series t Tp , center frequency f c and the time constant τ 1 , establish the duration model S Tp (Also called the duration pulse model).

[0099] By perturbation vector λ M , Duration Model S Tp and the starting time t of the partial discharge signal 0 , establish the time domain model S of the partial discharge signal 0m (t Tp ).

[0100] Specifically, the time series t of the partial discharge signal is calculated by formula (1): Tp .

[0101] (1);

[0102] In formula (1), T p is the pulse width, T s is the sampling interval; Function representation: uniformly generate 0 to T p The common "T p / T s "Point vector sequence, 0 is the starting point of the sequence, T p Equivalent to the last data, T p / T s Equivalent to the number of elements in the sequence.

[0103] The duration model S is established by formula (2)Tp .

[0104] (2);

[0105] In formula (2), A is the amplitude, τ 1 is the time constant, t Tp is a time series, τ is a variable, and f is a frequency.

[0106] The time domain model S of the PD signal is established by formula (3): 0m (t Tp ).

[0107] (3);

[0108] In formula (3), t 0 is the starting time of the partial discharge signal, t Tp is the time series, λ M is the disturbance vector, S Tp is a continuous time model, M is the number of partial discharge signals; m=1,2,…,M.

[0109] It should be noted that the duration model S shown in formula (2) Tp is the generation formula of a single partial discharge signal (equivalent to a single partial discharge pulse signal). Formula (3) represents the generation process of the mth partial discharge signal, that is, the disturbance vector λ M The mth value of S is multiplied by a single Tp , formula (3) can be used to simulate the generation of different single pulse duration signals.

[0110] Step S202: Based on the time domain model of the partial discharge signal, the signal number and Gaussian white noise, combined with fast Fourier transform, the frequency spectrum characteristics of the partial discharge signal are determined.

[0111] In the specific implementation of step S202, the number of partial discharge signals M is used to calculate the full time series t of the partial discharge signals. The full time series t is specifically: the total time length is M*T p , the sampling interval is T s sequence.

[0112] Time domain model S of joint partial discharge signal 0m (t Tp ) and Gaussian white noise (denoted as n(t)), the continuous partial discharge signal S(t) can be expressed as formula (4).

[0113] (4);

[0114] In formula (4), and They respectively represent the partial discharge signal of the first single pulse continuous time and the partial discharge signal of the second single pulse continuous time. The others are similar and will not be elaborated here.

[0115] Based on the continuous partial discharge signal S(t) shown in the above formula (4), the spectrum characteristic S of the continuous partial discharge signal can be obtained by fast Fourier transform. 0 (f), i.e. S 0 (f)=fft(S(t)).

[0116] Step S203: establishing a normalized spectrum distribution model of the partial discharge signal based on the spectrum characteristics of the partial discharge signal and in combination with the amplitude peak of the spectrum characteristics of the partial discharge signal.

[0117] In the specific implementation of step S203, based on the spectrum feature S of the partial discharge signal 0 (f) The peak amplitude A of the spectral characteristics of the partial discharge signal m , establish the normalized spectrum distribution model S(f) of the partial discharge signal, the normalized spectrum distribution model S(f) is shown in formula (5).

[0118] (5);

[0119] The above embodiments of the present invention Figure 2 , which is a description of how to establish a normalized spectrum distribution model S(f) of a partial discharge signal. By establishing the normalized spectrum distribution model S(f) of a partial discharge signal, it is possible to provide original signal support for subsequent frequency domain parameter extraction and interference suppression of the partial discharge signal.

[0120] For the above-mentioned embodiment of the present invention Figure 1 The spectrum distribution interval and the first spectrum characteristic parameter involved in step S103 are shown in Figure 3 , shows a flow chart of determining a spectrum distribution interval and a first spectrum characteristic parameter provided by an embodiment of the present invention, Figure 3 The steps include:

[0121] Step S301: using a normalized spectrum distribution model and threshold parameters, estimating the start frequency and cutoff frequency of the partial discharge signal.

[0122] In the specific implementation of step S301, based on the normalized spectrum distribution model S(f) of the continuous partial discharge signal, the threshold parameter a is input, so as to estimate the starting frequency (f) of the partial discharge signal. star ) and the cut-off frequency (f end ).

[0123] Specifically, the threshold parameter a is input into the normalized spectrum distribution model S(f) to estimate the start frequency and the cutoff frequency.

[0124] Among them, the starting frequency f star ={f | f > a}, the cut-off frequency f end ={f | f < a}, where f is the frequency.

[0125] Step S302: Determine the spectral distribution range of the partial discharge signal based on the frequency perturbation coefficient, starting frequency, and cut-off frequency.

[0126] In the specific implementation process of step S302, based on the frequency perturbation coefficient ξ, starting frequency f star and cut-off frequency f end , determine the spectral distribution range (denoted as freq) of the continuous partial discharge signal.

[0127] Specifically, the spectral distribution range freq = [f star *ξ, f end *ξ].

[0128] It should be noted that since the bandwidths of different pulses may vary, but the spectral distributions of partial discharge signals from the same source are basically the same, setting the frequency perturbation coefficient can provide redundancy for pulse width differences.

[0129] Step S303: Determine the first spectral characteristic parameter of the partial discharge signal based on the normalized spectral distribution model, starting frequency, and cut-off frequency.

[0130] In the specific implementation process of step S303, based on the normalized spectral distribution model, starting frequency, and cut-off frequency, determine the bandwidth and mean frequency component of the spectrum of the partial discharge signal; use the bandwidth and mean frequency component to determine the root mean square error, skewness, and steepness of the pulse waveform of the partial discharge signal, and the determined root mean square error, skewness, and steepness of the pulse waveform are the first spectral characteristic parameters of the partial discharge signal.

[0131] Specifically, based on the normalized spectral distribution model S(f), search for the extreme value distribution between the starting frequency f star and the cut-off frequency f end , solve the extreme value distribution curve S extre (f), and the specific content of the extreme value distribution curve S extre (f) can be seen in formula (6).

[0132] (6);

[0133] In formula (6), τ is a variable, f is the frequency, and dτ represents the calculus of the variable τ from positive infinity to negative infinity in the interval.

[0134] Based on the extreme value distribution curve S extre (f), combined with the starting frequency f starand the cutoff frequency f end The estimated result is used to determine the bandwidth of the spectrum of the mth partial discharge signal (denoted as B m ), that is, B m =f end -f star ; and determine the mean frequency component of the spectrum of the mth PD signal (denoted as ).

[0135] (7);

[0136] In formula (7), the length() function is used to calculate the total number of data in the sequence, and f is the frequency;

[0137] Meaning: According to the frequency from the starting frequency f star Initially, the frequency coordinate increases by one value each time until the frequency value reaches the cutoff frequency f end End, a sequence of S values ​​at this frequency sampling point.

[0138] Using the mean frequency component of the spectrum of the mth partial discharge signal and bandwidth B m , the RMS error of the pulse waveform of the mth PD signal (denoted as ).

[0139] (8);

[0140] Using the mean frequency component of the spectrum of the mth partial discharge signal , bandwidth B m and pulse waveform root mean square error , the skewness of the mth PD signal (denoted as p) is determined by formula (9) and formula (10) respectively. m ) and steepness (denoted as z m ).

[0141] (9);

[0142] (10);

[0143] In formula (9) and formula (10), f i is any frequency between the start frequency and the cut-off frequency, is the variance (RMS error of the pulse waveform ) to the power of 1.5, is the variance (RMS error of the pulse waveform ) squared.

[0144] Through the above formulas (8) to (10), the first spectrum characteristic parameter of the partial discharge signal can be determined. The skewness in the first spectrum characteristic parameter is recorded as p s , the steepness of the first spectral characteristic parameter is recorded as z s , the pulse waveform root mean square error in the first spectrum characteristic parameter is recorded as .

[0145] The above embodiments of the present invention Figure 3 , which is the relevant explanation on determining the spectrum distribution range and the first spectrum characteristic parameters. By determining the spectrum distribution range and the first spectrum characteristic parameters of the local discharge signal, the frequency domain analysis parameters can be determined, laying a parameter foundation for interference suppression of local discharge signals.

[0146] For the above-mentioned embodiment of the present invention Figure 1 For details on filtering the partial discharge signal in step S104, see Figure 4 , shows a flow chart of filtering a partial discharge signal provided by an embodiment of the present invention, Figure 4 The steps include:

[0147] Step S401: using the specified data, respectively establishing a co-frequency interference signal duration model and a variable frequency interference signal duration model.

[0148] In the specific implementation of step S401, the pulse width T of the partial discharge signal is used. p , amplitude A, time series t Tp , center frequency f c (Here the center frequency of the co-channel interference signal is represented by f c ) and the time constant τ 2 , establish the co-frequency interference signal duration model S Tpin1 , co-channel interference signal duration model S Tpin1 The specific content of is shown in formula (11).

[0149] (11);

[0150] Using the pulse width T of the partial discharge signal p , amplitude A, time series t Tp , center frequency f c1 (Here the center frequency of the variable frequency interference signal is expressed as f c1 ) and the time constant τ 3 , establish the variable frequency interference signal duration model S Tpin2 , variable frequency interference signal duration model S Tpin2 The specific content of is shown in formula (12).

[0151] (12);

[0152] The above is about the duration model S of the same-frequency interference signal Tpin1 And the variable frequency interference signal duration model S Tpin2 Description.

[0153] Step S402: establishing a mixed signal model based on the co-frequency interference signal duration model, the variable frequency interference signal duration model and the continuous partial discharge signal.

[0154] In the specific implementation of step S402, based on the co-frequency interference signal duration model S Tpin1 , variable frequency interference signal duration model S Tpin2 and the continuous partial discharge signal (that is, S(t) shown in the above formula (4)), as well as the co-channel interference signal duration model S Tpin1 , variable frequency interference signal duration model S Tpin2 Input constant L, build mixed signal model S all (t), mixed signal model S all The specific content of (t) is shown in formula (13).

[0155] (13);

[0156] In formula (13), is the first single pulse continuous time partial discharge signal, is the first single pulse interference signal with the same frequency and duration, is the first variable frequency single pulse duration interference signal, is the Lth single pulse interference signal with the same frequency, is the Lth frequency-converted single pulse duration interference signal, is the partial discharge signal of the Mth single pulse continuous time, and n(t) is Gaussian white noise.

[0157] Step S403: Perform fast Fourier transform on the mixed signal model to obtain the frequency spectrum characteristics of the mixed signal model.

[0158] In the specific implementation of step S403, the mixed signal model S all (t) Perform fast Fourier transform to obtain the mixed signal model S all The spectrum characteristics S of (t) all (f).

[0159] Step S404: Filter the frequency spectrum characteristics of the mixed signal model using a bandpass filter set by the frequency spectrum distribution interval to obtain a filtered partial discharge signal.

[0160] In the specific implementation of step S404, a bandpass filter (denoted as filter) is set based on the spectrum distribution interval (freq) of the continuous partial discharge signal, and the spectrum characteristics of the mixed signal model are filtered using the bandpass filter (filter) to obtain a filtered partial discharge signal S fliter (f), namely the filtered partial discharge signal S fliter (f) = S all (f)*filter.

[0161] The above embodiments of the present invention Figure 4 , which is a description of filtering partial discharge signals.

[0162] After obtaining the filtered partial discharge signal, a second spectrum characteristic parameter is determined for the filtered partial discharge signal. The method for determining the second spectrum characteristic parameter can refer to the above-mentioned embodiment of the present invention. Figure 3 The relevant contents of determining the first spectrum characteristic parameters in will not be repeated here.

[0163] The first frequency spectrum characteristic parameter of the partial discharge signal is a characteristic parameter determined for a pure partial discharge signal, and the second frequency spectrum characteristic parameter of the filtered partial discharge signal is a characteristic parameter determined after interference suppression.

[0164] According to the above embodiments of the present invention Figure 1 As can be seen from the content, the skewness in the first spectrum characteristic parameter is recorded as p s , the steepness of the first spectral characteristic parameter is recorded as z s , the pulse waveform root mean square error in the first spectrum characteristic parameter is recorded as The skewness in the second spectrum characteristic parameter is recorded as p all , the steepness of the second spectrum characteristic parameter is recorded as z all , the pulse waveform root mean square error in the second spectrum characteristic parameter is recorded as .

[0165] Determine the parameter difference (Δp, ΔS) between the second spectrum characteristic parameter and the first spectrum characteristic parameter 2 and Δz), where Δp=p all -p s , ΔS 2 = - , Δz=z all -z s .

[0166] Input error ratio x, the parameter difference of the signal released by the authority (Δp, ΔS 2 and Δz) satisfy “{Δp, ΔS 2, Δz}>x”, the partial discharge signal is eliminated. Otherwise, the parameter difference of the partial discharge signal (Δp, ΔS 2 and Δz) do not satisfy “{Δp, ΔS 2 , Δz}>x”, the partial discharge signal is retained.

[0167] The parameter difference (Δp, ΔS) between the second spectrum characteristic parameter and the first spectrum characteristic parameter is 2 and Δz), eliminating the conditions, the interference suppression result S of the partial discharge signal can be extracted R (t), interference suppression result S R The PD signal that has completed interference suppression in (t) is the retained PD signal, and the PD signal that has been removed is equivalent to the interference signal. In this way, interference signals with different frequency interval distributions and different spectral characteristics are suppressed, and interference signal suppression processing can be achieved during PD signal detection.

[0168] In order to verify the effectiveness of the interference suppression technology proposed in this scheme, the effectiveness of the interference suppression technology of frequency domain intelligent filtering was verified by building a real prototype test model. The main simulation parameters used in the verification process are: time constant τ 1 is 5ns, the duration of the partial discharge signal or interference signal is T p is 150ns, the time interval of partial discharge is 500ns, the carrier frequency of partial discharge signal is 0.18GHz and 0.41GHz, the carrier frequency of interference signal is 0.62GHz, the sampling frequency is 1GHz, the gate width is 80 / sampling frequency, the starting time of partial discharge signal is t 0 =1.5*duration time*sampling frequency, the half-order of the variable scale difference filter is 85, the accumulation coefficient and subtraction coefficient of the variable scale difference filter are 1, and the sampling interval is 1.

[0169] After verifying the effectiveness of the interference suppression technology of frequency domain intelligent filtering, it can be obtained Figures 5 to 8 The data shown, where Figures 5 to 8 In the diagram, the units of the coordinate axes include amplitude, frequency, pulse numbers, and time.

[0170] in, Figure 5 This is an example of the spectrum characteristics of a continuous partial discharge signal. Figure 5 As shown in the example diagram of the frequency spectrum characteristics, it can be seen that the partial discharge signal has two frequency components.

[0171] Figure 6 This is an example of the spectrum characteristics of the interference signal. Figure 6It can be seen from the example diagram of the frequency spectrum characteristics of the interference signal that the interference signal and the partial discharge signal are not in the same frequency range after the frequency is set.

[0172] Figure 7 This is an example of the three-dimensional distribution of the partial discharge signal and the interference signal after time sequence alignment. The X-axis is time, the Y-axis is the number of pulses, and the Z-axis is the amplitude. Figure 7 It can be seen that the three-dimensional distribution simulates the aliasing of local discharge signals and interference signals in the acquisition process, and the interference signals cannot be eliminated and determined only from the time domain.

[0173] Figure 8 This is an example of the three-dimensional distribution of the interference suppression of continuous partial discharge signals using this scheme, that is, effective suppression of interference signals is achieved through frequency domain parameter comparison. Figure 8 The gaps in the three-dimensional distribution example diagram are the result of interference signals being suppressed.

[0174] Corresponding to the method for extracting partial discharge signals based on frequency domain intelligent filtering provided by the above-mentioned embodiment of the present invention, see Fig. 9 , the embodiment of the present invention also provides a structural block diagram of a partial discharge signal extraction system based on frequency domain intelligent filtering, the partial discharge signal extraction system includes: an acquisition unit 100, an establishment unit 200, a first determination unit 300, a filtering unit 400, a second determination unit 500 and a removal unit 600;

[0175] The acquisition unit 100 is used to acquire the specified data of the continuous partial discharge signal of the high-voltage electrical equipment, and the specified data at least includes: pulse width, sampling interval, center frequency, signal number and amplitude.

[0176] The establishing unit 200 is used to establish a normalized spectrum distribution model of the partial discharge signal by using the specified data.

[0177] The first determination unit 300 is used to determine the spectrum distribution interval and the first spectrum characteristic parameter of the partial discharge signal based on the normalized spectrum distribution model.

[0178] The filtering unit 400 is used to filter the partial discharge signal using the spectrum distribution interval and the specified data.

[0179] The second determining unit 500 is used to determine a second frequency spectrum characteristic parameter of the filtered partial discharge signal.

[0180] The elimination unit 600 is used to eliminate the partial discharge signal based on the first spectrum characteristic parameter and the second spectrum characteristic parameter to extract the partial discharge signal with interference suppression completed.

[0181] In the embodiments of the present invention, spectral feature parameters are used for frequency-domain feature comparison and analysis, so as to suppress interference signals during the detection process of partial discharge signals, and finally extract the partial discharge signals with interference suppression completed, improving the accuracy and reliability of detecting partial discharge signals.

[0182] Preferably, in combination with Fig. 9 the content shown, the establishment unit 200 includes a first establishment module, a determination module, and a second establishment module, and the execution principles of each module are as follows:

[0183] The first establishment module is used to establish a time-domain model of the partial discharge signal by using the pulse width, sampling interval, and center frequency.

[0184] The determination module is used to determine the spectral features of the partial discharge signal based on the time-domain model of the partial discharge signal, the number of signals, and Gaussian white noise, in combination with the fast Fourier transform.

[0185] The second establishment module is used to establish a normalized spectral distribution model of the partial discharge signal based on the spectral features of the partial discharge signal, in combination with the amplitude peak value of the spectral features of the partial discharge signal.

[0186] Preferably, in combination with Fig. 9 the content shown, the first determination unit 300 includes an estimation module, a first determination module, and a second determination module, and the execution principles of each module are as follows:

[0187] The estimation module is used to estimate the start frequency and cut-off frequency of the partial discharge signal by using the normalized spectral distribution model and the threshold parameter.

[0188] The first determination module is used to determine the spectral distribution interval of the partial discharge signal based on the frequency perturbation coefficient, the start frequency, and the cut-off frequency.

[0189] The second determination module is used to determine the first spectral feature parameter of the partial discharge signal based on the normalized spectral distribution model, the start frequency, and the cut-off frequency.

[0190] In some embodiments, the first spectral feature parameter at least includes skewness, kurtosis, and root mean square error of the pulse waveform; specifically, the second determination module is used for:

[0191] Based on the normalized spectral distribution model, the start frequency, and the cut-off frequency, determine the bandwidth and mean frequency component of the spectrum of the partial discharge signal;

[0192] Use the bandwidth and mean frequency component to determine the root mean square error, skewness, and kurtosis of the pulse waveform of the partial discharge signal.

[0193] Preferably, in combination with Fig. 9The content shown, the filtering unit 400 includes a first establishment module, a second establishment module, a transformation module and a filtering module, and the execution principle of each module is as follows:

[0194] The first establishing module is used to establish a co-frequency interference signal duration model and a variable frequency interference signal duration model respectively by using designated data.

[0195] The second establishing module is used to establish a mixed signal model based on the same-frequency interference signal duration model, the variable-frequency interference signal duration model and the continuous partial discharge signal.

[0196] The transformation module is used to perform fast Fourier transform on the mixed signal model to obtain the frequency spectrum characteristics of the mixed signal model.

[0197] The filtering module is used to filter the frequency spectrum characteristics of the mixed signal model by using a bandpass filter set by the frequency spectrum distribution interval to obtain a filtered partial discharge signal.

[0198] Preferably, combined Fig. 9 The content shown, the elimination unit 600 includes a calculation module and a elimination module, and the execution principle of each module is as follows:

[0199] The calculation module is used to calculate the parameter difference between the first spectrum characteristic parameter and the second spectrum characteristic parameter.

[0200] The elimination module is used to eliminate the partial discharge signal for each partial discharge signal if the parameter difference corresponding to the partial discharge signal meets the elimination condition; if the parameter difference corresponding to the partial discharge signal does not meet the elimination condition, determine that the partial discharge signal is a partial discharge signal that completes interference suppression.

[0201] In summary, the embodiments of the present invention provide a method and system for extracting partial discharge signals based on frequency domain intelligent filtering, which utilizes frequency spectrum characteristic parameters to perform frequency domain feature comparison and analysis, thereby suppressing interference signals in the partial discharge signal detection process, and finally extracting partial discharge signals with interference suppression, thereby improving the accuracy and reliability of detecting partial discharge signals.

[0202] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without creative work.

[0203] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0204] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for extracting partial discharge signals based on frequency domain intelligent filtering, characterized in that: The method comprises: Acquire specified data of a continuous partial discharge signal of a high-voltage electrical device, wherein the specified data includes at least: pulse width, sampling interval, center frequency, number of signals and amplitude; Using the specified data, establishing a normalized spectrum distribution model of the partial discharge signal; Based on the normalized spectrum distribution model, determining the spectrum distribution interval and the first spectrum characteristic parameter of the partial discharge signal; Filtering the partial discharge signal using the frequency spectrum distribution interval and the designated data; Determining a second frequency spectrum characteristic parameter of the filtered partial discharge signal; Based on the first frequency spectrum characteristic parameter and the second frequency spectrum characteristic parameter, the partial discharge signal is eliminated to extract the partial discharge signal with interference suppression completed.

2. The method according to claim 1, characterized in that Using the specified data, a normalized spectrum distribution model of the partial discharge signal is established, including: Establishing a time domain model of the partial discharge signal by using the pulse width, the sampling interval and the center frequency; Based on the time domain model of the partial discharge signal, the signal number and Gaussian white noise, combined with fast Fourier transform, determining the frequency spectrum characteristics of the partial discharge signal; Based on the frequency spectrum characteristics of the partial discharge signal and in combination with the amplitude peak value of the frequency spectrum characteristics of the partial discharge signal, a normalized frequency spectrum distribution model of the partial discharge signal is established.

3. The method according to claim 1, characterized in that Determining the spectrum distribution interval and the first spectrum characteristic parameter of the partial discharge signal based on the normalized spectrum distribution model includes: Using the normalized spectrum distribution model and threshold parameters, estimating the starting frequency and the cutoff frequency of the partial discharge signal; Determining a frequency spectrum distribution interval of the partial discharge signal based on a frequency disturbance coefficient, the starting frequency and the cut-off frequency; Based on the normalized spectrum distribution model, the start frequency and the cutoff frequency, a first spectrum characteristic parameter of the partial discharge signal is determined.

4. The method according to claim 3, characterized in that The first frequency spectrum characteristic parameters include at least skewness, steepness and pulse waveform root mean square error; Determining a first spectrum characteristic parameter of the partial discharge signal based on the normalized spectrum distribution model, the start frequency, and the cutoff frequency includes: Determining a bandwidth and a mean frequency component of the spectrum of the partial discharge signal based on the normalized spectrum distribution model, the start frequency and the cutoff frequency; The frequency bandwidth and the mean frequency component are used to determine the root mean square error, skewness and steepness of the pulse waveform of the partial discharge signal.

5. The method according to claim 1, characterized in that Filtering the partial discharge signal using the frequency spectrum distribution interval and the designated data includes: Using the specified data, respectively establish a same-frequency interference signal duration model and a variable-frequency interference signal duration model; Establishing a mixed signal model based on the co-frequency interference signal duration model, the variable frequency interference signal duration model and the continuous partial discharge signal; Performing a fast Fourier transform on the mixed signal model to obtain a frequency spectrum feature of the mixed signal model; The frequency spectrum characteristics of the mixed signal model are filtered using a bandpass filter set by the frequency spectrum distribution interval to obtain the filtered partial discharge signal.

6. The method according to any one of claims 1 to 5, characterized in that: Based on the first spectrum characteristic parameter and the second spectrum characteristic parameter, the partial discharge signal is eliminated to extract the partial discharge signal for completing interference suppression, including: Calculating a parameter difference between the first frequency spectrum characteristic parameter and the second frequency spectrum characteristic parameter; For each of the partial discharge signals, if the parameter difference corresponding to the partial discharge signal satisfies a rejection condition, the partial discharge signal is rejected; If the parameter difference corresponding to the partial discharge signal does not satisfy the elimination condition, it is determined that the partial discharge signal is the partial discharge signal that completes interference suppression.

7. A partial discharge signal extraction system based on frequency domain intelligent filtering, characterized in that: The system comprises: An acquisition unit, used to acquire specified data of a continuous partial discharge signal of a high-voltage electrical device, wherein the specified data at least includes: pulse width, sampling interval, center frequency, number of signals and amplitude; An establishing unit, used to establish a normalized spectrum distribution model of the partial discharge signal using the specified data; A first determining unit, configured to determine a spectrum distribution interval and a first spectrum characteristic parameter of the partial discharge signal based on the normalized spectrum distribution model; A filtering unit, configured to filter the partial discharge signal using the frequency spectrum distribution interval and the designated data; A second determining unit, used to determine a second frequency spectrum characteristic parameter of the filtered partial discharge signal; The elimination unit is used to eliminate the partial discharge signal based on the first spectrum characteristic parameter and the second spectrum characteristic parameter to extract the partial discharge signal with interference suppression completed.

8. The system according to claim 7, characterized in that The establishing unit comprises: A first establishing module, used to establish a time domain model of the partial discharge signal by using the pulse width, the sampling interval and the center frequency; A determination module, used to determine the frequency spectrum characteristics of the partial discharge signal based on the time domain model of the partial discharge signal, the signal number and Gaussian white noise in combination with fast Fourier transform; The second establishing module is used to establish a normalized spectrum distribution model of the partial discharge signal based on the spectrum characteristics of the partial discharge signal and in combination with the amplitude peak value of the spectrum characteristics of the partial discharge signal.

9. The system according to claim 7, characterized in that The first determining unit includes: An estimation module, used to estimate the starting frequency and the cutoff frequency of the partial discharge signal by using the normalized spectrum distribution model and the threshold parameter; A first determination module, configured to determine a frequency spectrum distribution interval of the partial discharge signal based on a frequency disturbance coefficient, the start frequency and the cutoff frequency; The second determination module is used to determine a first spectrum characteristic parameter of the partial discharge signal based on the normalized spectrum distribution model, the start frequency and the cutoff frequency.

10. The system according to claim 9, characterized in that The first spectrum characteristic parameter includes at least skewness, steepness and pulse waveform root mean square error; the second determination module is specifically used for: Determining a bandwidth and a mean frequency component of the spectrum of the partial discharge signal based on the normalized spectrum distribution model, the start frequency and the cutoff frequency; The frequency bandwidth and the mean frequency component are used to determine the root mean square error, skewness and steepness of the pulse waveform of the partial discharge signal.

Images