Multi-quantity output relay protection secondary circuit test method and system

Abstract

The application relates to the technical field of smart grids, and discloses a multi-electricity output relay protection secondary circuit test method and system, wherein the multi-electricity output relay protection secondary circuit test method comprises the following steps: collecting an environmental noise time sequence, obtaining a noise spectrum and setting a filter parameter set; outputting a test signal comprising a characteristic code signal based on the noise spectrum and the filter parameter set; executing a plurality of test sequences using the test signal to obtain a segmented signal; performing real-time noise suppression on the segmented signal to obtain an enhanced signal; extracting a feature vector from the enhanced signal; and obtaining a fault type and a fault type position coordinate based on the feature vector; through noise spectrum analysis and independent component analysis, the application realizes high anti-interference capability and real-time noise suppression of the test signal, successfully separates a composite fault and identifies the fault type, and through cross-correlation functions and a coordinate optimization algorithm, accurate positioning of the fault is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of smart grid, more particularly, it relates to a multi-electricity output relay protection secondary circuit test method and system. BACKGROUND

[0002] As an important guarantee for the safe and stable operation of the power system, the reliability of the secondary circuit of the relay protection is directly related to the normal work of the relay protection device.

[0003] The existing Chinese patent with publication number CN110888082A discloses a relay protection secondary circuit node voltage fault positioning method and device, which includes constructing a secondary circuit model. The secondary circuit model includes a positive power supply node, a negative power supply node, a plurality of elements located between the positive power supply node and the negative power supply node, and a plurality of nodes at both ends of the elements. The voltage of each node in the secondary circuit model is obtained to form a node voltage matrix. A normal state voltage model is provided, which includes a plurality of normal voltage matrices, and the fault point is located according to the comparison between the node voltage matrix and the normal state voltage model. This method realizes secondary circuit fault diagnosis and fault positioning by collecting electrical quantities of the secondary circuit, and further realizes real-time fault monitoring of the secondary circuit, thereby improving the operation and maintenance management level of the smart substation.

[0004] However, in the actual test environment, electromagnetic interference, background noise and the like can seriously affect the quality of the test signal, resulting in the true value of the multiple fault signals generated by the composite faults commonly existing in the relay protection secondary circuit being distorted. The existing test method lacks adaptive anti-interference ability in complex electromagnetic environments, and it is difficult to effectively separate and identify the multiple fault signals mixed together, and it is even more difficult to accurately locate the specific position of the fault occurrence, resulting in low maintenance efficiency. SUMMARY

[0005] The purpose of the present application is to provide a multi-electricity output relay protection secondary circuit test method and system to solve the above problems.

[0006] The present application provides a multi-electricity output relay protection secondary circuit test method, comprising the following steps:

[0007] Collecting an environmental noise time series, mapping the time series to a frequency domain, obtaining a noise spectrum and setting a filter parameter set;

[0008] Based on the noise spectrum and the filter parameter set, outputting a test signal including a characteristic code signal;

[0009] Using the generated test signal to perform a plurality of test sequences, collecting response data of the measured circuit under different excitation modes, and performing segmented processing on the response data to obtain segmented signals;

[0010] Real-time noise suppression is performed on the segmented signal to obtain an enhanced signal;

[0011] According to the enhanced signal and the characteristic code signal, a signal component containing various fault components is separated from the aliasing enhanced signal, and a feature vector capable of representing fault characteristics is extracted according to the signal component;

[0012] Based on the feature vector, a hierarchical fault recognition strategy is adopted to obtain a fault type and a fault type position coordinate.

[0013] Further, the method for obtaining a noise spectrum and setting a filter parameter set comprises:

[0014] Step 101, sampling and analyzing the background noise of the test environment to obtain a noise time sequence; using Fourier transform to map the noise time sequence to the frequency domain to obtain a noise spectrum;

[0015] Step 102, determining the main noise frequency band and its corresponding energy distribution according to the noise spectrum analysis, the main noise frequency band being a frequency band with energy exceeding a preset threshold; for each noise frequency band, setting filter parameters based on center frequency, bandwidth parameter, stopband attenuation, and passband fluctuation;

[0016] Step 103, combining the filter parameters of the multiple noise frequency bands and an adaptive adjustment factor into a filter parameter set, wherein the adaptive adjustment factor is used to dynamically adjust the filter parameters according to noise changes during the test.

[0017] Further, the method for outputting a test signal comprises:

[0018] Step 201, applying a dynamic phase compensation algorithm to introduce real-time phase calculation for each channel, the real-time phase being the sum of an initial phase, a real-time frequency integral, and a dynamic compensation calculated based on the filter parameter set;

[0019] Step 202, calculating the dynamic compensation in the form of a weighted sum of multiple frequency components, which is realized by superposition of multiple frequency components, each frequency component including a compensation coefficient, a phase adjustment function, and a sine waveform;

[0020] Step 203, setting a harmonic superposition formula with noise adaptive adjustment, the harmonic superposition formula calculating a test signal for each channel, the test signal for each channel being the sum from the fundamental wave to the high-order harmonic, and each harmonic component being the product of a basic amplitude, a noise suppression adjustment factor, and a sine wave plus a characteristic code signal;

[0021] Step 204, using the weighted sum of the multi-frequency sinusoidal components as the characteristic code signal; the characteristic code signal is the sum of low-amplitude coefficients multiplied by a plurality of sinusoidal components, each sinusoidal component having a specific amplitude, a characteristic frequency and an initial phase, wherein the specific amplitude refers to the relative strength value preset for each characteristic code component.

[0022] Further, the specific method for obtaining the segmented signal is as follows:

[0023] Step 301, converting the test signal into various excitation forms;

[0024] Step 302, for each excitation mode, collecting the response data of the circuit under test, the response data including voltage data, current data, phase data and frequency data;

[0025] Step 303, using a time window function to segment the circuit response data to obtain the segmented signal.

[0026] Further, the method for obtaining the enhanced signal is:

[0027] Step 401, using an adaptive Kalman filter to perform real-time noise suppression on the segmented signal to obtain a Kalman-filtered signal;

[0028] Step 402, using wavelet packet decomposition to further denoise the Kalman-filtered signal, expressing the Kalman-filtered signal as a linear combination of wavelet basis functions of different scales and shifts and wavelet coefficients, and obtaining processed wavelet coefficients by threshold processing on the wavelet coefficients;

[0029] Step 403, reconstructing the time-domain signal by inverse wavelet transform, mapping the processed wavelet coefficients back to the time domain to obtain the enhanced signal.

[0030] Further, the specific method for extracting a feature vector capable of representing fault characteristics according to the signal components is:

[0031] Step 501, applying a blind signal separation algorithm to separate the overlapped enhanced signal to obtain a separated signal component;

[0032] Step 502, calculating the cross-correlation coefficients of each separated signal component and the characteristic code signal, and when the cross-correlation coefficient is greater than a set threshold, considering that the component contains effective fault information, and setting the separated signal component as an effective signal component;

[0033] Step 503, extracting a multi-dimensional feature containing time-domain features, frequency-domain features and time-frequency domain features from the effective signal component;

[0034] Step 504, combine the multi-dimensional features to form a high-dimensional feature vector, the high-dimensional feature vector contains time domain features, frequency domain features and time-frequency domain features, apply principal component analysis for feature dimension reduction to obtain a reduced feature vector, the reduced feature vector will be output as a feature vector.

[0035] Further, the method for separating the mixed enhanced signal by applying blind signal separation algorithm to obtain the separated signal components comprises:

[0036] The enhanced signal is regarded as an observation signal matrix, various independent fault signals are regarded as an independent fault source signal matrix, and a coefficient matrix describing how the fault signals are mixed into the observation signal is regarded as a mixing matrix, a linear model between the observation signal matrix and the independent fault source signal matrix is established, the observation signal matrix in the linear model is a linear combination of the multiple independent fault source signal matrices and the mixing matrix, wherein the mixing matrix determines the contribution degree of each fault source to the observation signal;

[0037] A fast independent component analysis algorithm is used to solve the separation matrix, the separation matrix is an inverse matrix or a pseudo-inverse matrix of the mixing matrix, the mixed observation signal is converted back to the independent fault signal through the separation matrix to obtain an estimated source signal matrix, each row in the estimated source signal matrix represents a separated signal component.

[0038] Further, the specific method for obtaining the fault type and the fault type position coordinates comprises:

[0039] Step 601, a three-level classifier architecture is adopted to refine the fault type and obtain a hierarchical classification result;

[0040] Step 602, the relevance of the hierarchical classification result is inferred using a Bayesian network to determine the causal relationship and co-occurrence probability between multiple faults, and multiple simultaneous fault types are obtained;

[0041] Step 603, for each fault type, the signal propagation time difference is used for positioning, and the distance from the fault type to the test point is obtained according to the product of the signal propagation speed in the medium and the time difference of the fault signal reaching the test point divided by 2.

[0042] Step 604, the distance information of multiple test points is used to obtain the position coordinates of the fault type by minimizing the fault coordinate optimization function.

[0043] Further, the time difference is determined by a cross-correlation function, the cross-correlation function measures the change of the similarity of two signals with time offset, when the offset makes the two signals most matched, the cross-correlation function reaches the maximum value, and the offset is the time difference; the cross-correlation function calculates the correlation between the enhanced signal related to the fault type and the test point reference signal, and finds the time delay value that makes the correlation maximum as the time difference.

[0044] The application provides a multi-power output relay protection secondary circuit test system for storing computer readable instructions, which can execute the aforementioned multi-power output relay protection secondary circuit test method when the computer readable instructions are read.

[0045] The application has the beneficial effects that: the application realizes high anti-interference capability of the test signal through noise spectrum analysis and adaptive filter parameter setting; the dynamic phase compensation algorithm is adopted to ensure the synchronization accuracy of the multi-channel signal; the quality of signal acquisition is improved through the feature code embedding and multi-band segmented acquisition technology; the real-time noise suppression is realized through the adaptive Kalman filtering and wavelet packet decomposition combination method; the independent component analysis and Bayesian network inference are applied to successfully separate the composite fault and identify the fault type; and the fault is accurately positioned through the cross-correlation function and coordinate optimization algorithm. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1 is a flowchart of a multi-power output relay protection secondary circuit test method of the application;

[0047] Figure 2 is a method example diagram of the application for extracting a feature vector capable of representing fault characteristics according to a signal component;

[0048] Figure 3 is a method example diagram of the application for obtaining a fault type and a fault type position coordinate;

[0049] Figure 4 is a module example diagram of a multi-power output relay protection secondary circuit test system of the application. DETAILED DESCRIPTION

[0050] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that discussions of these implementations are merely provided to enable those skilled in the art to better understand and thus implement the subject matter described herein, and are not intended to limit the scope of the disclosure. Various modifications can be made to the function and arrangement of elements discussed without departing from the scope of the subject matter described herein. Each example can omit, substitute or add various procedures or components also in accordance with the application. Additionally, features described in relation to some examples can also be combined in other examples.

[0051] A multi-power output relay protection secondary circuit test method and system, including the following embodiments:

[0052] Embodiment 1

[0053] A multi-power output relay protection secondary circuit test method, as shown in Figure 1 includes the following steps:

[0054] Step 100, collect the environmental noise time series, map the time series to the frequency domain, obtain the noise spectrum and set the filter parameter set.

[0055] The method of obtaining the noise spectrum and setting the filter parameter set comprises:

[0056] Step 101, sampling and analyzing the background noise of the test environment to obtain the noise time series; in order to convert the time domain noise signal in the noise time series to the frequency domain, Fourier transform is used to map the noise time series to the frequency domain to obtain the noise spectrum.

[0057] Step 102, according to the noise spectrum analysis, determine the main noise frequency band and its corresponding energy distribution, the specific method is: first, calculate the energy density of the noise spectrum, identify the frequency band whose energy exceeds the preset threshold (2 times of the average energy), and take these frequency bands as the main noise frequency band; then apply the peak detection algorithm to locate the local maximum points of the energy density, and take these points as the center frequency of the main noise frequency band; then calculate the energy distribution characteristics of each noise frequency band, including the frequency band width (the frequency range where the energy decreases to half of the peak value) and the energy concentration (the ratio of the peak energy to the total energy of the frequency band).

[0058] For the Ni-th noise frequency band, set the filter parameter θ Ni of the Ni-th noise frequency band based on the center frequency, bandwidth parameter, stopband attenuation, passband fluctuation

[0059] θ Ni ={f c,Ni ,BW Ni ,A s,Ni ,δ Ni}

[0060] Where, θ Ni is the filter parameter of the Ni-th noise frequency band, f c,Ni is the center frequency or cutoff frequency of the Ni-th noise frequency band (determines the center position of the frequency band where the filter acts), BW Ni is the bandwidth parameter of the Ni-th noise frequency band (controls the frequency range covered by the filter, which needs to cover the noise frequency band width), A s,Ni is the stopband attenuation of the Ni-th noise frequency band (determines the suppression strength outside the noise frequency band, which needs to meet the noise suppression requirement), and δ Ni is the passband fluctuation of the Ni-th noise frequency band (limits the allowable distortion of signals in the passband of the filter, and guarantees the quality of useful signals).

[0061] The center frequency, bandwidth parameters, stopband attenuation, and passband ripple are determined by minimizing the objective function, which is the integral of the product of the filter transfer function and the noise spectrum. The optimization objective is to find the filter parameter combination that most effectively suppresses a specific noise frequency band. Specifically, the process involves finding the filter parameter combination that most effectively suppresses a specific noise frequency band, with the center frequency f of the noise band as the starting point. Noise,Ni With this as the core, within the bandwidth (±Δf), the center frequency, bandwidth parameters, stopband attenuation, and passband ripple of the filter are optimized to minimize the integral of the product of the filter's transfer function and the noise spectrum.

[0062] For the filter parameter θ of the Ni-th noise frequency band Ni The specific calculation method for optimizing the objective function is as follows:

[0063]

[0064] Where H(f,θ) Ni Let f be the transfer function of the filter at frequency f, and Δf be the bandwidth considered. Noise,Ni N represents the center of the noise frequency band in the Ni-th noise frequency band. NS (f) represents the noise spectrum obtained after Fourier transform. The transfer function is determined by the filter parameters θ of the Ni-th noise frequency band. Ni Completely determined; for example, for a bandpass filter, its transfer function H(f,θ) Ni There is a direct relationship between the filter and its parameters; the filter parameters determine the shape and characteristics of the transfer function. Specifically, the center frequency determines the frequency range at which the filter operates, the bandwidth parameter controls the passband width, the stopband attenuation is reflected by the filter order and determines the degree to which the filter suppresses stopband signals, and the passband ripple affects the fidelity of the signal within the passband. These parameters together constitute the complete characteristics of a bandpass filter, enabling it to effectively suppress noise in a specific frequency band while retaining useful signals. The specific calculation formula is as follows:

[0065]

[0066] Where n is the filter order, and is determined by the required stopband attenuation A of the Ni-th noise frequency band. s,Ni The higher the order, the steeper the transition band and the greater the stopband attenuation.

[0067] The center frequency, bandwidth parameter, stopband attenuation, and passband ripple parameter are optimized using a gradient descent method combined with a grid search strategy. First, the center frequency f... c,Ni Initialized to the center frequency of the noise band, bandwidth parameter BW Ni Initialize the bandwidth to the noise band, then perform an iterative search in the parameter space. In each iteration, calculate the gradient of the objective function with respect to each parameter, and update the parameter values ​​in the reverse direction of the gradient. Stopband attenuation As,Ni By adjusting the filter order n, the higher the order, the greater the attenuation; passband fluctuation δ Ni Then by introducing a normalization coefficient control in the transfer function, the coefficient is directly related to the gain at the passband edge.

[0068] Step 103, combine the filter parameters of multiple noise bands and the adaptive adjustment factor into a filter parameter set, and the adaptive adjustment factor is used to dynamically adjust the filter parameters according to the noise changes during the test process.

[0069] Wherein, the setting logic of the adaptive adjustment factor is based on the difference degree between the real-time noise spectrum and the initial environmental noise spectrum, by calculating the correlation coefficient and energy ratio between the two, when the noise characteristics are detected to change significantly, the system will automatically increase the adjustment factor value to speed up the filter parameter update speed, so that the filtering system can quickly adapt to the new noise environment, and ensure the quality and reliability of the test signal. The correlation coefficient refers to the statistical similarity between the real-time noise spectrum and the initial environmental noise spectrum, which is obtained by calculating the ratio of the covariance and standard deviation of the two spectra at each frequency point, the value range is [-1, 1], the closer to 1, the more similar the two spectra; the energy ratio is the ratio of the total energy of the real-time noise spectrum and the total energy of the initial environmental noise spectrum, which reflects the overall change of the noise intensity, when the ratio deviates from 1 significantly, it means that the noise environment has changed significantly, and the filter parameters need to be updated.

[0070] Step 200, based on the noise spectrum and the filter parameter set, output a test signal including a characteristic code signal.

[0071] The method for outputting an anti-interference test signal comprises:

[0072] Step 201, in order to ensure the synchronization of the multi-channel signal and dynamically adapt to the noise environment, a dynamic phase compensation algorithm is applied, a real-time phase is introduced for each channel, the real-time phase is the sum of the initial phase, the real-time frequency integral and the dynamic compensation amount calculated based on the filter parameter set; for the i-th channel, its real-time phase θ i (t) is:

[0073]

[0074] Wherein, θ i,0 is the initial phase (to ensure the initial synchronization of the multi-channel signal), is the real-time frequency integral (to track the change of frequency with time, ensure the continuity and dynamic synchronization of the phase), f(t) is the real-time frequency at time t, represents the integral from time 0 to time t, dt represents the integral variable, Δθ i (t, Θ Filter ) is the dynamic compensation amount calculated based on the filter parameter set ΘFilter The calculated dynamic compensation amount (adapted to the noise environment based on the filter parameter set, compensating for the phase shift caused by noise).

[0075] In this embodiment, a channel refers to an independent path for transmitting a specific type of electrical signal. Specifically, a channel can be divided into a voltage channel and a current channel, each corresponding to a physical output port for providing an analog signal to the measured relay protection device. For example, in a three-phase power system, there are usually three voltage channels (corresponding to A, B, and C three-phase voltages) and three current channels (corresponding to A, B, and C three-phase currents), as well as possible zero sequence or neutral point channels. Each channel can independently control the amplitude, phase, frequency, and other parameters of its output signal, but needs to maintain precise synchronization with other channels to simulate the operating state of the real power system.

[0076] Step 202, in order to dynamically adjust the phase compensation according to the noise spectrum characteristics, the dynamic compensation amount is calculated in the form of weighted sum of multiple frequency components. Each frequency component is composed of three parts: compensation coefficient, phase adjustment function, and sinusoidal waveform. The compensation coefficient is used to control the weight of each frequency component; the phase adjustment function dynamically adjusts the compensation amount according to the noise intensity at the corresponding frequency; the sinusoidal waveform provides periodic compensation matching the noise frequency. Select the center frequency of 3-5 main noise frequency bands as the compensation frequency point, when the noise intensity at these frequency points increases, the corresponding compensation amount will also increase, thereby achieving targeted phase compensation and effectively improving the anti-interference ability of the test signal in complex electromagnetic environment. Its calculation formula is:

[0077]

[0078] where α k is the compensation coefficient (used to weight the compensation intensity of different noise frequency bands, its specific calculation method is to multiply the ratio of the energy of the kth noise frequency band to the total energy of all selected noise frequency bands by the maximum compensation coefficient (usually 0.2)), φ k is the phase adjustment function (dynamically calculates the compensation amount according to the noise spectrum characteristics), K is the number of compensation terms (covers multiple frequency bands to improve robustness, the value range is [3, 5], covers the main noise frequency band to improve the anti-interference ability of the system), f k is the compensation frequency point (directed compensation for frequency bands with significant noise, the center frequency of the main noise frequency band), sin(2πf k t) is the sinusoidal waveform, which provides periodic compensation matching the noise frequency.

[0079] φ k (N NS (f k))The phase adjustment function converts the noise spectrum intensity into a phase adjustment amount using a nonlinear mapping function, ensuring that a small amount of compensation is provided when the noise is weak, and a larger compensation is provided when the noise is strong, but does not exceed the safety threshold. The nonlinear mapping function is: taking the smaller one of 0.5 and the product value, which is 0.1 times the logarithmic value of: 1 plus 10 times the noise spectrum intensity at frequency f divided by the logarithmic value of the reference noise level, where the reference noise level is usually taken as the average value of the ambient noise. k

[0080] Step 203, in order to generate a test signal containing multiple harmonics and capable of avoiding noise frequency bands, a harmonic superposition formula with noise adaptive adjustment is set, which calculates the test signal of the i-th channel by summing from the first harmonic to the 24th harmonic. Each harmonic component is composed of the product of the basic amplitude, the noise suppression adjustment factor and the sine wave, plus the characteristic code signal. Among them, the noise suppression adjustment factor is dynamically adjusted according to the noise spectrum intensity at the harmonic frequency, so that the harmonic energy automatically avoids the frequency band where the noise is concentrated; the sine wave is generated according to the real-time phase; the characteristic code signal is used for subsequent signal identification, and the specific calculation method is:

[0081]

[0082] where V i (t) is the test signal of the i-th channel, with units of volts (V) or amperes (A), depending on the channel type; represents the sum from the first harmonic to the 24th harmonic, n is the harmonic index, A n is the basic amplitude of the n-th harmonic, with the same unit as V i (t); β is the noise suppression coefficient (determined according to the sensitivity of the system to noise and the importance of the test signal, with a value range of 0.5 to 2.0, in a conventional test scenario, β takes a medium value (such as 0.8-1.2), providing balanced noise suppression effect; in a scenario where signal integrity is the priority, β takes a smaller value (such as 0.5-0.7), only moderately suppressing strong noise frequency bands, and automatically increasing the β value to strengthen the suppression effect when detecting a deteriorating noise environment); N NS (nh) represents the noise spectrum at frequency nh, (1-β·N NS (nh)) is the noise suppression adjustment factor, which is used to dynamically adjust the actual amplitude of each harmonic, so that the harmonic energy avoids the frequency band where the noise is concentrated, and is a dimensionless quantity; h is the fundamental frequency, sin(n×θ i (t)) is a sine wave with a value range of [-1, 1] and is dimensionless; S Mark (t) is a characteristic code signal with the same unit as V i (t), ensuring that it can be directly added to the harmonic component.​

[0083] Base amplitude A n is multiplied by a noise suppression adjustment factor (1-β·N NS (nh)) and a sinusoidal wave sin(n×θ i (t)), resulting in a product with the same unit as A n , i.e. volts or amperes; the signature signal S Mark (t) also has the same unit, so they can be directly added to form the final test signal V i (t) for the i-th channel. The test signals V i (t) for multiple channels constitute the complete test signal V(P).

[0084] Step 204, in order to embed a low-amplitude signature in the signal for subsequent identification of valid signals, a weighted sum of multiple frequency sinusoidal components is used as the signature signal. The signature signal is composed of a low-amplitude coefficient (with a value range of 0.05-0.1) multiplied by the sum of multiple sinusoidal components, each with a specific amplitude, characteristic frequency, and initial phase, where the specific amplitude refers to the relative strength value pre-set for each signature component, which is optimized according to signal identification requirements and anti-interference requirements, and is usually allocated according to a specific pattern (such as a decreasing sequence or the golden ratio) to ensure the uniqueness and recognizability of the signature. The characteristic frequency is selected in the noise-minimum frequency band identified in step 100 to ensure that the signature remains identifiable without being disturbed by noise. The specific calculation method is as follows:

[0085]

[0086] where γ is the low-amplitude coefficient (with a value range of 0.05-0.1), and P is the number of signature components (determined according to the balance between the accuracy of signature identification and the complexity of calculation required by the system. In a standard test environment, P is usually set to 3-5, which can provide sufficient complexity of the signature to ensure reliable identification while maintaining computational efficiency; in a high-interference environment, P can be increased to 6-8 to improve anti-interference ability by increasing the redundancy of the signature; in scenarios requiring fast processing, P can be reduced to 2-3 to reduce computational burden. In addition, the value of P is also related to the uniqueness required by the signature, when the system needs to distinguish more different types of test signals, a larger P value should be used to provide more coding space. The characteristic frequency f m of each signature component is uniformly distributed within the noise-minimum frequency band to maximize the spectral utilization efficiency and identification reliability of the signature, B m is the amplitude of the m-th component, f m is the characteristic frequency of the m-th component (selected in the noise-minimum frequency band identified in step 100), is the initial phase of the m-th component.

[0087] Step 300, using the generated test signal to perform a plurality of test sequences, collect the response data of the measured circuit under different excitation modes, and segment the response data to obtain segmented signals. The response data includes voltage, current, phase and frequency data.

[0088] The specific method for obtaining the segmented signal is as follows:

[0089] Step 301, in order to comprehensively test the measured circuit using different excitation modes, the basic test signal is converted into various excitation forms, including the following modes:

[0090] Steady state mode: directly output the original test signal without any transformation processing, used for testing the response characteristics of the measured circuit under steady state;

[0091] Step mode: suddenly apply the test signal at the step time, simulate the mutation in the power system, wherein the step time refers to the time point when the signal suddenly changes from zero to the set value;

[0092] Sweep mode: make the frequency of the test signal continuously change within a certain range, gradually change from the starting frequency to the terminal frequency, and the whole change process is completed within a sweep cycle, used for testing the response characteristics of the measured circuit under different frequencies;

[0093] Pulse mode: convert the test signal into a series of periodic pulse sequences, the time interval between each pulse is the pulse period, and a certain number of pulses are generated in total, used for testing the response characteristics of the measured circuit to transient signals.

[0094] Step 302, for each excitation mode, collect the response data of the measured circuit, including voltage data, current data, phase data and frequency data:

[0095] Voltage data: A, B, C three-phase and neutral point voltage;

[0096] Current data: A, B, C three-phase current;

[0097] Phase data: phase of each voltage and current;

[0098] Frequency data: A, B, C three-phase frequency change.

[0099] Step 303, in order to reduce spectral leakage and improve the analysis accuracy of different frequency bands, a multi-band segmentation collection method is adopted, and a time window function w j (t) is used to segment the circuit response data X(t) to obtain the segmented signal, and the specific expression formula is:

[0100] X Se,j(t) = X(t) x w j (t)

[0101] where the time window function w j (t) includes three types of rectangular windows, Hanning windows, and Blackman windows; different segments cover low, medium, and high frequency bands, each frequency band uses different sampling rates and window lengths, and different window functions are used for processing to reduce spectral leakage and interference effects.

[0102] Different segments cover different frequency bands, and one example of specific division is as follows:

[0103] Low frequency band: f ∈ [10, 100] Hz, rectangular window is used, sampling rate f s = 2 kHz, window length T w = 200 ms;

[0104] Medium frequency band: f ∈ [100, 500] Hz, Hanning window is used, sampling rate f s = 5 kHz, window length T w = 100 ms;

[0105] High frequency band: f ∈ [500, 1200] Hz, Blackman window is used, sampling rate f s = 10 kHz, window length T w = 50 ms.

[0106] Step 400: Real-time noise suppression is performed on the segmented signal to obtain an enhanced signal.

[0107] The specific method for obtaining the enhanced signal is:

[0108] Step 401, using an adaptive Kalman filter to perform real-time noise suppression on the segmented signal to obtain a Kalman filtered signal.

[0109] Step 402, although Kalman filtering can effectively suppress Gaussian noise, but the effect is limited for non-Gaussian noise and transient interference. Wavelet packet decomposition is used to further denoise the Kalman filtered signal, which is expressed as a linear combination of wavelet basis functions of different scales and translations and wavelet coefficients, and the specific expression is:

[0110]

[0111] where ∑ j,p represents the sum of all scales j and translations p, c j,p is the wavelet coefficient, and ψ j,p(t) is a wavelet basis function; a wavelet basis function is a special function with compact support (local finite), oscillation and attenuation, and through scaling and translation operations, it can form a complete function family, which can effectively represent the local time-frequency characteristics of a signal; and a wavelet coefficient is a projection value of a signal on a corresponding wavelet basis function, which reflects the energy distribution of the signal at a specific time scale and position, and a larger coefficient usually corresponds to an important feature in the signal, and a smaller coefficient may represent a noise component.

[0112] In order to remove the noise component in the wavelet domain, a hard threshold processing is applied to the wavelet coefficients to obtain processed wavelet coefficients, and coefficients greater than a threshold are retained while coefficients less than the threshold are suppressed, so that the noise component can be effectively removed while the important features of the signal are retained, and the specific calculation method is:

[0113]

[0114] wherein, represents the processed wavelet coefficient, c j,p represents the original wavelet coefficient, λ j represents the threshold value of the jth scale.

[0115] In order to determine the optimal threshold value according to the signal length and the noise level, a general threshold value calculation formula based on the noise standard deviation is adopted, the threshold value of the jth scale is the product of the noise standard deviation and the optimal threshold value factor, the optimal threshold value factor is the natural logarithm of 2 times the signal length N, and the specific calculation formula is:

[0116]

[0117] wherein, σ j is the noise standard deviation of the jth scale, which is calculated by the energy distribution of the noise spectrum in the corresponding frequency band (the specific calculation method is to take the square root of the energy value of the noise spectrum in the frequency band corresponding to the jth scale), N is the signal length (the number of sampling points in the signal), is the optimal threshold value factor derived based on statistical theory, lnN is the natural logarithm of the signal length N, and the threshold value is appropriately increased with the increase of the signal length.

[0118] For segmented signals of different frequency bands, different wavelet basis functions are adopted:

[0119] Low frequency band (10-100Hz): Daubechies-4 wavelet is used;

[0120] Medium frequency band (100-500Hz): Daubechies-8 wavelet is used;

[0121] High frequency band (500-1200Hz): Symlet-6 wavelet is used.

[0122] Step 403, finally, the time domain signal is reconstructed by inverse wavelet transform, the processed wavelet coefficients are mapped back to the time domain to obtain the enhanced signal; the enhanced signal is the sum of the product of all wavelet coefficients processed by threshold and the corresponding wavelet basis function, and the specific calculation method is as follows:

[0123]

[0124] Wherein, X En (t) is the enhanced signal.

[0125] Step 500: According to the enhanced signal and the characteristic code signal, the signal components containing various fault components are separated from the mixed enhanced signal, and the feature vectors capable of representing the fault characteristics are extracted according to the signal components.

[0126] The method for extracting feature vectors capable of representing fault characteristics according to signal components is as shown in Figure 2 , and specifically includes:

[0127] Step 501, in the relay protection secondary circuit, multiple faults often exist at the same time and are superimposed on each other to form a composite fault signal, and the enhanced signal X En (t) is used to separate the mixed enhanced signal by applying a blind signal separation algorithm to obtain the separated signal components.

[0128] The method for obtaining the separated signal components includes:

[0129] The enhanced signal X En (t) is regarded as an observation signal matrix X; various independent fault signals are regarded as an independent fault source signal matrix S; and the coefficient matrix describing how the fault signals are mixed into the observation signal is regarded as a mixing matrix A.

[0130] In order to represent the linear mixing relationship of multiple fault signals, a linear model is established between the observation signal matrix and the independent fault source signal matrix. The observed signal in the linear model is a linear combination of multiple independent fault signals, wherein the mixing matrix determines the contribution degree of each fault source to the observation signal, and the specific expression formula is as follows:

[0131] X=AS

[0132] In order to recover the original independent fault signal from the mixed signal, a separation matrix is calculated; the separation matrix is the inverse matrix or pseudo-inverse matrix of the mixing matrix, and the mixed observation signal is converted back to the independent fault signal through the separation matrix to obtain an estimated source signal matrix, each row in the estimated source signal matrix represents a separated signal component, and the conversion from the mixed observation signal to the independent fault component is realized by applying the separation matrix to the observation signal matrix. The specific expression formula is as follows:

[0133]

[0134] wherein, is the estimated source signal matrix, i.e. the independent fault signal matrix after separation, W represents the separation matrix, and X represents the observed signal matrix.

[0135] The method for calculating the separation matrix comprises:

[0136] The separation matrix W is solved using a Fast Independent Component Analysis algorithm (FastICA), which maximizes a non-Gaussianity measure through an iterative optimization process. The non-Gaussianity measure is an index for measuring the degree of deviation of the mixed signal from the normal distribution (Gaussian distribution), and the iterative optimization process maximization means that the algorithm repeatedly calculates multiple times to constantly adjust the parameters to maximize the non-Gaussianity index.

[0137] In order to quantify the non-Gaussianity of the signal and serve as a measure of independence, a contrast function based on a nonlinear function is defined, which calculates the square of the difference between the nonlinear function expectation value of the input signal vector (i.e. the observed mixed signal) and the standard Gaussian random variable (i.e. the normal distribution random variable with a mean of 0 and a variance of 1). The nonlinear function is selected as a function such as tanh or exp, which is used to enhance the non-Gaussianity of the signal; the contrast function is used as the optimization target, and the greater the value, the better the separation effect; the difference in expectation value reflects the degree of difference between the signal distribution and the Gaussian distribution.

[0138] Step 502, after completing the signal separation, it is necessary to identify which separated components contain effective fault information; by calculating the cross-correlation coefficient of each signal component after separation and the characteristic code signal, when the cross-correlation coefficient is greater than a set threshold (the value range is 0.6-0.8), it is considered that the component contains effective fault information, and the separated signal component is set as an effective signal component.

[0139] The cross-correlation coefficient calculation method is to divide the sum of the product of the separated signal component and the characteristic code after subtracting their mean values by the product of the standard deviations of the two. The higher the cross-correlation coefficient, the more likely it is that the component contains effective signal information; the specific calculation method of the cross-correlation coefficient is:

[0140]

[0141] wherein, ρ i is the cross-correlation coefficient, s i (t) is the i-th separated signal component, is the mean value of the signal component s i (t), is the mean value of the characteristic code signal S Mark (t), and Σ t represents the sum over all time points t. The cross-correlation coefficient ρ iThe higher, the more likely the component contains valid signal information.

[0142] Step 503, after determining the valid signal component, extract multi-dimensional features from the valid signal component, feature extraction includes the following three types of features:

[0143] Time domain features: obtained by directly statistical analysis of time domain signal, including mean, variance, peak value, kurtosis, skewness, etc.

[0144] Frequency domain features: obtained by Fourier transform of the signal in the frequency domain analysis, including power spectral density, harmonic content, frequency band energy distribution, etc.

[0145] Time-frequency domain features: obtained by analyzing the characteristics of the signal in time and frequency domain, including wavelet energy entropy, empirical mode decomposition (EMD) coefficient, etc. These features can reflect the frequency characteristics of the signal over time.

[0146] Step 504, in order to fully characterize the fault characteristics, the above extracted different types of features are combined to form a high-dimensional feature vector, which contains time domain features, frequency domain features and time-frequency domain features. However, the high-dimensional feature vector may have redundant information, affecting the efficiency and accuracy of subsequent fault diagnosis. Therefore, principal component analysis (PCA) is applied for feature dimension reduction, and the reduced feature vector is obtained by multiplying the transpose of the principal component transformation matrix with the original feature vector. The reduced feature vector will be output as the feature vector.

[0147] Step 600: Based on the feature vector, a hierarchical fault recognition strategy is adopted to obtain the fault type and fault type position coordinates.

[0148] The method for obtaining the fault type and fault type position coordinates is as shown in Figure 3 , specifically including:

[0149] Step 601, a three-level classifier architecture is adopted to gradually refine the fault type and obtain a hierarchical classification result.

[0150] In the first level classification, the system first processes the reduced feature vector to realize the coarse classification of fault type. The first level classifier receives the reduced feature vector F Dr as input, and outputs the main type of fault C1 according to the preset classifier parameters, such as ground fault, short circuit fault or open circuit fault, etc.

[0151] In the second level classification, the system further subdivides the fault subtypes based on the coarse classification. The second level classifier not only considers the original feature vector information, but also takes the result C1 of the first level classification as an important reference basis, and outputs a more detailed fault subtype classification result C2 by comprehensively analyzing the two parts of information.

[0152] In the third-level classification, the system performs a final, precise classification of the fault. The three-level classifier comprehensively considers the feature vectors and the classification results C1 and C2 from the first two levels. Through more complex classification algorithms and parameter configurations, it determines the specific fault mode and severity C3, providing detailed guidance for subsequent fault handling.

[0153] This hierarchical classification architecture can gradually improve the accuracy of classification. Through this progressive classification method, complex fault conditions can be identified more accurately, and a complete hierarchical classification result (C1, C2, C3) can be obtained.

[0154] For example, consider a fault signal in the secondary circuit of a relay protection system:

[0155] The first-level classifier identifies it as a "short circuit fault" (C1);

[0156] The second-level classifier further subdivides it into "phase-to-phase short circuits" (C2);

[0157] The third-level classifier ultimately determined it to be "a metallic short circuit between phase A and phase B, with a high severity" (C3).

[0158] Step 602: After obtaining the hierarchical classification results (C1, C2, C3), a Bayesian network is used to infer the correlation between the hierarchical classification results, determine the causal relationship and co-occurrence probability among multiple faults, and obtain multiple simultaneously existing fault types.

[0159] The core of Bayesian network inference is to associate observed evidence with failure probabilities to calculate the posterior probability, which is the probability that a certain failure exists given the observation of specific evidence E. This probability equals the probability of observing evidence E when the failure exists (likelihood probability) multiplied by the failure F. i The prior probability is divided by the total probability of observed evidence E under all possible faults (as a normalization factor). Observed evidence E consists of two parts: feature vector and hierarchical classification results (C1, C2, C3).

[0160] The formula for Bayes' theorem is:

[0161]

[0162] Among them, P(F) i |E) represents the condition that evidence E is observed, and the fault F is... i The probability of existence (posterior probability), P(E|F) i ) indicates fault F i When evidence E exists, the probability of observing it (likelihood probability) is P(F). i ) indicates fault Fi the prior probability, i.e. the base probability of failure occurrence without any observation evidence, and the denominator is the normalization factor to ensure the sum of all failure probabilities is 1.

[0163] The prior probability P(F i ) is obtained by statistical analysis of historical failure data, reflecting the occurrence frequency of different failures in historical records. Bayesian networks store the dependency between nodes through conditional probability tables, enabling the system to handle uncertainty and causality in compound failures and ultimately identify multiple simultaneous failure types

[0164] Step 603, time difference analysis of fault signals. For each identified failure type F i , determine its specific location in the secondary circuit. Use the signal propagation time difference for positioning, and the product of the signal propagation speed in the medium and the time difference of the fault signal arriving at the test point divided by 2 to obtain the distance of the failure type F i to the test point l.

[0165] The formula for calculating the distance of the fault to the test point is:

[0166]

[0167] where d il represents the distance of the failure type F i to the test point l, v represents the signal propagation speed in the medium, determined by the material properties of the circuit, Δt il represents the time difference of the fault signal arriving at the test point, and dividing by 2 is because of the characteristics of signal round-trip propagation.

[0168] The time difference Δt il is determined by the cross-correlation function, which measures the similarity of two signals as the time offset changes. When the offset makes the two signals most compatible, the cross-correlation function reaches a maximum value, and this offset is the time difference; specifically, the cross-correlation function calculates the correlation between the enhanced signal related to the failure type and the reference signal at the test point, and finds the time delay value that maximizes the correlation as the time difference. The calculation method is:

[0169]

[0170] where, is the cross-correlation function of the failure type F i , and is the enhanced signal related to the failure type F i . is the reference signal of test point l (the reference signal is obtained by collecting the signal at test point j under normal working state at initialization moment, and is band-pass filtered and normalized to improve signal quality, and the reference signal is used as a benchmark for comparison with the fault signal), argmax τ denotes the time delay τ that makes the cross-correlation function reach the maximum value, denotes the integral from negative infinity to positive infinity.

[0171] In step 604, the position coordinates of the fault type are obtained by minimizing the fault coordinate optimization function using the distance information of multiple test points. The optimization objective of the fault coordinate optimization function is to find a spatial coordinate point, so that the sum of squares of errors between the theoretical distance from the point to each test point and the actual distance calculated by time difference analysis is minimized. This is a minimization problem, which finds the coordinate point that minimizes the sum of squares of errors between the estimated distance (fault to test point distance calculation) and the actual distance (i.e. the Euclidean distance between the coordinate point and the test point). The square operation ensures that positive and negative errors are treated equally, and large errors are amplified.

[0172] The formula of the fault coordinate optimization function is:

[0173]

[0174] where (x i ,y il ,z i ) is the spatial coordinate of the fault type F (x,y,z) , d l denotes the distance from the fault type F l to the test point l, argmin l denotes the coordinate (x, y, z) that makes the following expression reach the minimum value, the expression calculates the sum of squares of errors between the estimated distance and the actual distance, the square root term calculates the Euclidean distance between the coordinate point and the test point, and the square operation ensures that positive and negative errors are treated equally, and large errors are amplified, denotes the summation from l = 1 to l = R, (x j ,y j ,z j ) is the coordinate of the lth test point, and R is the number of test points. This minimization problem is solved by the Levenberg-Marquardt algorithm, which iteratively updates the coordinate estimate.

[0175] In the relay protection secondary circuit, the coordinate system of the spatial coordinate is defined as a three-dimensional rectangular coordinate system, in which:

[0176] X-axis: along the width direction of the power distribution cabinet or control cabinet, unit: meter (m);

[0177] Y-axis: along the depth direction of the power distribution cabinet or control cabinet, unit: meter (m);

[0178] Z-axis: along the height direction of the power distribution cabinet or control cabinet, unit: meter (m);

[0179] The coordinate origin (0, 0, 0) is set at the lower left corner of the front end of the power distribution cabinet or control cabinet. The test point coordinates (x j ,y j ,z j ) are fixed positions measured and recorded in the system in advance, and these test points are usually distributed at key nodes of the relay protection secondary circuit, such as terminal blocks, relays, transducers, etc.

[0180] In order to improve the positioning accuracy, the following principles should be followed in the arrangement of test points:

[0181] The number of test points R should be no less than 4 to ensure a unique solution for three-dimensional space positioning;

[0182] Test points should be distributed in different planes as much as possible to avoid collinear or coplanar arrangement;

[0183] The test point spacing should be appropriate, neither too dense nor too scattered, and the spacing is usually 0.5m to 2m.

[0184] Through this coordinate positioning method, the system can accurately determine the spatial position of the fault, and the positioning accuracy is usually ±10cm, which provides intuitive fault location information for maintenance personnel and greatly shortens the fault troubleshooting time.

[0185] Embodiment 2

[0186] Referring to Figure 4 , a multi-electricity output relay protection secondary circuit test system is provided, which is used to store computer readable instructions capable of executing the aforementioned multi-electricity output relay protection secondary circuit test method when the computer readable instructions are read, and the system comprises:

[0187] An environment perception module 101 acquires an environmental noise time series, maps the time series to the frequency domain, obtains a noise spectrum, and sets a filter parameter set;

[0188] A collaborative output module 102 outputs a test signal including a feature code signal based on the noise spectrum and the filter parameter set;

[0189] A test acquisition module 103 uses the generated test signal to perform a plurality of test sequences, acquires response data of the measured circuit under different excitation modes, and performs segmented processing on the response data to obtain segmented signals; wherein the response data includes voltage, current, phase and frequency data;

[0190] The enhancement module 104 performs real-time noise suppression on the segmented signal to obtain an enhanced signal;

[0191] The fault separation module 105 separates signal components containing various fault components from the mixed enhanced signal according to the enhanced signal and the characteristic code signal, and extracts a characteristic vector capable of representing fault characteristics according to the signal components;

[0192] The diagnosis and positioning module 106 obtains a fault type and a fault type position coordinate based on the characteristic vector and using a hierarchical fault recognition strategy.

[0193] The above describes the embodiments of the present application, but the embodiments are not limited to the specific implementation described above, which is only illustrative but not restrictive. Those skilled in the art can make more equivalent embodiments under the inspiration of the embodiments, which are all within the protection scope of the embodiments.

Claims

1. A multi-energy output relay protection secondary circuit test method, characterized in that, The method comprises the following steps: Collecting an environmental noise time sequence, mapping the time sequence to a frequency domain, obtaining a noise spectrum, and setting a filter parameter set; Based on the noise spectrum and the filter parameter set, output a test signal including a characteristic code signal; Using the generated test signal to perform a plurality of test sequences, collecting response data of the measured loop under different excitation modes, and performing segmented processing on the response data to obtain segmented signals; Performing real-time noise suppression on the segmented signals to obtain enhanced signals; According to the enhanced signals and the characteristic code signal, separating the mixed enhanced signals to obtain signal components containing various fault components, and extracting a feature vector capable of representing fault characteristics according to the signal components, the specific method being: Step 501, applying a blind signal separation algorithm to separate the mixed enhanced signals to obtain separated signal components; Step 502, calculating the cross-correlation coefficients of the separated signal components and the characteristic code signal, when the cross-correlation coefficient is greater than a set threshold, it is considered that the component contains effective fault information, and the separated signal component is set as an effective signal component; Step 503, extracting a multi-dimensional feature containing a time domain feature, a frequency domain feature and a time-frequency domain feature from the effective signal component; Step 504, combining the multi-dimensional feature to form a high-dimensional feature vector, the high-dimensional feature vector contains a time domain feature, a frequency domain feature and a time-frequency domain feature, applying principal component analysis for feature dimension reduction to obtain a reduced feature vector, and the reduced feature vector will be output as a feature vector; Based on the feature vector, a hierarchical fault recognition strategy is adopted to obtain a fault type and a fault type position coordinate.

2. The multi-quantity output relay protection secondary circuit test method according to claim 1, characterized in that, The method for obtaining the noise spectrum and setting the filter parameter set comprises: Step 101, sampling and analyzing the background noise of the test environment to obtain a noise time sequence; using Fourier transform to map the noise time sequence to a frequency domain to obtain a noise spectrum; Step 102, determining the main noise frequency band and its corresponding energy distribution according to the noise spectrum analysis, the main noise frequency band being a frequency band with energy exceeding a preset threshold; for each noise frequency band, setting filter parameters based on the center frequency, bandwidth parameter, stopband attenuation and passband fluctuation; Step 103, combining the filter parameters of the plurality of noise frequency bands and an adaptive adjustment factor into a filter parameter set, wherein the adaptive adjustment factor is used to dynamically adjust the filter parameters according to noise changes during the test.

3. The multi-quantity output relay protection secondary circuit test method according to claim 1, characterized in that, The method for outputting the test signal comprises: Step 201, applying a dynamic phase compensation algorithm to introduce real-time phase calculation for each channel, the real-time phase being the sum of an initial phase, real-time frequency integration and a dynamic compensation calculated based on the filter parameter set; Step 202, calculating the dynamic compensation in the form of weighted sum of multiple frequency components, which is realized by superposition of multiple frequency components, each frequency component including a compensation coefficient, a phase adjustment function and a sine wave; Step 203, setting a harmonic superposition formula with noise adaptive adjustment, the harmonic superposition formula calculating a test signal of each channel, the test signal of each channel being the sum from a fundamental wave to a high-order harmonic, and each harmonic component being the product of a basic amplitude, a noise suppression adjustment factor and a sine wave plus a characteristic code signal; Step 204, using the weighted sum of multiple frequency sinusoidal components as the feature code signal; the feature code signal is the sum of multiple sinusoidal components multiplied by low amplitude coefficients, each sinusoidal component has a specific amplitude, a characteristic frequency and an initial phase, wherein the specific amplitude refers to the relative strength value preset for each feature code component.

4. The multi-quantity output relay protection secondary circuit test method according to claim 1, characterized in that, The specific method for obtaining the segmented signal is as follows: Step 301, converting the test signal into various excitation forms; Step 302, for each excitation mode, collecting the response data of the circuit under test, the response data including voltage data, current data, phase data and frequency data; Step 303, using a time window function to segment the circuit response data to obtain the segmented signal.

5. The multi-quantity output relay protection secondary circuit test method according to claim 1, characterized in that, The method for obtaining the enhanced signal is: Step 401, using an adaptive Kalman filter to perform real-time noise suppression on the segmented signal to obtain a Kalman filtered signal; Step 402, using wavelet packet decomposition to further denoise the Kalman filtered signal, expressing the Kalman filtered signal as a linear combination of wavelet basis functions of different scales and shifts and wavelet coefficients, and obtaining processed wavelet coefficients by threshold processing on the wavelet coefficients; Step 403, reconstructing the time domain signal by inverse wavelet transform, mapping the processed wavelet coefficients back to the time domain to obtain the enhanced signal.

6. The multi-quantity output relay protection secondary circuit test method according to claim 1, characterized in that, The method for applying a blind signal separation algorithm to separate the mixed enhanced signal to obtain the separated signal components includes: Regarding the enhanced signal as an observation signal matrix; regarding various independent fault signals as an independent fault source signal matrix; regarding a coefficient matrix describing how the fault signals are mixed into the observation signal as a mixing matrix, establishing a linear model between the observation signal matrix and the independent fault source signal matrix, wherein the observation signal matrix in the linear model is a linear combination of the independent fault source signal matrix and the mixing matrix, and the mixing matrix determines the contribution degree of each fault source to the observation signal; Using a fast independent component analysis algorithm to solve a separation matrix; the separation matrix is the inverse matrix or pseudo-inverse matrix of the mixing matrix, and the mixed observation signal is converted back to the independent fault signal through the separation matrix to obtain an estimated source signal matrix, and each row in the estimated source signal matrix represents a separated signal component.

7. The multi-quantity output relay protection secondary circuit test method according to claim 1, characterized in that, The specific method for obtaining the fault type and the fault type position coordinates includes: Step 601, using a three-level classifier architecture to refine the fault type to obtain a hierarchical classification result; Step 602, using a Bayesian network to infer the relevance of the hierarchical classification result to determine the causal relationship and co-occurrence probability between multiple faults, and obtaining multiple simultaneous fault types; Step 603, for each fault type, using signal propagation time difference for positioning, and obtaining the distance from the fault type to the test point according to the product of the signal propagation speed in the medium and the time difference of the fault signal arriving at the test point divided by 2; Step 604, using the distance information of multiple test points to obtain the position coordinates of the fault type by minimizing the fault coordinate optimization function.

8. The multi-quantity output relay protection secondary circuit test method according to claim 7, characterized in that, The time difference is determined by a cross-correlation function, which measures the similarity of two signals as a function of time offset, and reaches a maximum when the offset is such that the two signals match best, and this offset is the time difference; The cross-correlation function calculates the correlation between the enhanced signal related to the fault type and the reference signal of the test point, and finds the time delay value that makes the correlation maximum as the time difference.

9. A multi-quantity output relay protection secondary circuit test system, characterized in that, The computer readable instructions are used to store the computer readable instructions, which can execute the multi-electricity output relay protection secondary circuit test method according to any one of claims 1-8 when read. The computer readable instructions are used to store the computer readable instructions, which can execute the multi-electricity output relay protection secondary circuit test method according to any one of claims 1-8 when read.

Images