Harmonic detection method and system of vehicle network electrical system

Abstract

The invention provides a harmonic detection method and system for a vehicle network electrical system, and relates to the technical field of harmonic detection, and the harmonic detection method for the vehicle network electrical system comprises the steps: obtaining a current signal of a vehicle network bus, and carrying out the band limit processing of the current signal; constructing a variable window width S transformation kernel function with self-adaptive frequency; based on bispectrum analysis of a third-order cumulant, harmonic phase coupling characteristics are extracted, and an enhanced spectrum is generated through principal component projection; generating a dynamic detection threshold by using a non-harmonic interval background noise model; and carrying out harmonic parameter fitting by adopting an exponential decay model with frequency domain priori constraint. According to the method, through a cascade architecture of adaptive time-frequency analysis, high-order statistical feature extraction, dynamic threshold detection and prior constraint fitting, full-process closed-loop optimization from signal preprocessing to parameter optimization is realized, and the method is particularly suitable for accurate detection of heavy-frequency, nonlinear and time-varying harmonics of a vehicle network system; and a diagnosis basis with higher reliability can be provided for electric energy quality treatment.

Description

Technical Field

[0001] The present invention relates to the technical field of harmonic detection, and in particular to a harmonic detection method and system for a vehicle network electrical system. Background Art

[0002] With the rapid development of rail transit, the widespread use of power electronic equipment in vehicle-grid electrical systems (e.g., traction power supply systems) has led to an increasingly prominent problem of harmonic pollution. For example, high-frequency switching devices (such as IGBTs) and nonlinear loads (such as rectifiers and inverters) will generate complex broadband harmonics, which not only affect the quality of power, but may also cause equipment overheating, resonance, and even system instability.

[0003] In the existing related technologies, the current harmonic detection scheme has the following problems: First, the resolution of high-frequency harmonic detection is insufficient. The traditional Fourier transform (FFT) has poor adaptability to non-stationary signals, and the fixed time-frequency window of the short-time Fourier transform (STFT) is difficult to take into account both low-frequency resolution and high-frequency time positioning capabilities.

[0004] Second, it is difficult to extract nonlinear harmonic features. Harmonics generated by nonlinear loads have phase coupling characteristics (such as sum frequency / difference frequency components), and conventional power spectrum analysis cannot capture such high-order statistical features.

[0005] Third, the set threshold has poor adaptability. Existing methods mostly use fixed thresholds or static models based on empirical values, which are difficult to deal with time-varying background noise (such as pulse interference and random fluctuations) in the vehicle network system.

[0006] Fourth, parameter fitting errors accumulate. Traditional harmonic decomposition methods (such as the Prony algorithm) are easily affected by noise, resulting in parameter estimation errors. Summary of the invention

[0007] In order to at least solve some of the technical problems in the related art, the present invention provides a harmonic detection method and system for a vehicle network electrical system.

[0008] In order to achieve the above object, the technical solution adopted by the present invention includes: According to a first aspect of the present invention, a method for detecting harmonics in a vehicle-grid electrical system is provided, comprising the following steps: Step S1: obtaining a current signal of a busbar of a vehicle network and performing band-limiting processing on the current signal; Step S2: construct a frequency-adaptive variable-width S-transform kernel function; Step S3: extracting harmonic phase coupling features based on bispectral analysis of third-order cumulants, and generating enhanced spectrum through principal component projection; Step S4: Generate a dynamic detection threshold using a background noise model in a harmonic-free interval; Step S5: using an exponential decay model with frequency domain prior constraints to perform harmonic parameter fitting.

[0009] Optionally, the step S1 specifically includes: Step S1-1: using a broadband current sensor to collect the current signal of the vehicle network bus; Step S1-2: Band-limiting processing through anti-aliasing filter: In the formula, is the complex frequency domain transfer function of the filter, is a complex frequency variable, is the cut-off angular frequency, is the filter order.

[0010] Optionally, the cutoff frequency is set to ; The filter order is set to .

[0011] Optionally, the step S2 specifically includes: The frequency-adaptive variable-window-width S-transform kernel function is constructed according to the following formula: In the formula, is the time-frequency matrix, For the moment The instantaneous value of the current signal at time is the frequency adaptive window width adjustment function, , is the window width ratio coefficient, is a very small frequency offset, is the Fourier basis function.

[0012] Optionally, in step S2, it also includes: When detected When the frequency adaptive window width adjustment function Adjusted to: in, is the frequency judgment threshold.

[0013] Optionally, the step S3 specifically includes: Step S3-1: Calculate the third-order cumulant: Step S3-2: Construct a bispectral feature matrix: Step S3-3: Implement principal component projection: In the formula, is the third-order cumulant of the signal, is the mathematical expectation, and is the time delay, For the moment The instantaneous value of the current signal at time is the current signal at time point Shift right based on The observed value after sampling intervals, is the current signal at time point Shift right based on The observed value after sampling intervals, is the bispectral value, and are two independent frequency variables, is the projected spectrum.

[0014] Optionally, the step S4 specifically includes: Step S4-1: Establish background noise model: Step S4-2: Dynamic threshold calculation: In the formula, For frequency An estimate of the standard deviation of the background noise at , is the number of sampling points in the harmonic-free interval, is the time-frequency matrix at frequency ,time The amplitude at ∈ harmonic-free interval, For frequency The detection threshold at is the empirical coefficient, Set a value for the false alarm rate.

[0015] Optionally, in step S4, the empirical coefficient is set to , the false alarm rate setting value is set to .

[0016] Optionally, the step S5 specifically includes: Step S5-1: Use the following algorithm to perform parameter fitting: Step S5-2: Optimize the constraints as follows: In the formula, For the The amplitude of the subharmonics, is the attenuation factor, For the The subharmonic frequencies, is the initial phase, is the model order, is the number of sampling points, is the index value of the sampling point, , is the sampling interval, The current signal reconstructed by the algorithm is The predicted value of the sampling point, is the regularization coefficient, According to the a priori estimated value of the harmonic frequency obtained in steps S2 to S4.

[0017] According to a second aspect of the present invention, a harmonic detection system of a vehicle grid electrical system is further provided, which is used to perform the harmonic detection method of the vehicle grid electrical system as described in any technical solution in the first aspect of the present invention, and the harmonic detection system of the vehicle grid electrical system includes: The data acquisition and processing module is used to obtain the current signal of the vehicle network bus and perform band-limiting processing on the current signal; Adaptive time-frequency analysis module, used to construct frequency-adaptive variable-width S-transform kernel function; Bispectral feature extraction module, used for bispectral analysis based on third-order cumulants, extracting harmonic phase coupling features, and generating enhanced spectrum through principal component projection; A dynamic threshold generation module, used to generate a dynamic detection threshold using a harmonic-free interval background noise model; The parameter optimization module is used to fit the harmonic parameters using an exponential decay model with frequency domain prior constraints.

[0018] Beneficial effects: 1. Through the above technical solution, first, by performing band-limiting processing on the current signal, not only can the Nyquist sampling theorem be effectively ensured to hold, but also high-frequency noise aliasing can be effectively suppressed, that is, the interference of high-frequency noise on harmonic detection can be suppressed.

[0019] Second, by constructing a frequency-adaptive variable-width S-transform kernel function, it can be well adapted to signals in different frequency bands. For low-frequency signals, the variable-width S-transform kernel function can effectively ensure its high frequency resolution. For high-frequency signals, the frequency domain leakage caused by time window widening can be suppressed through exponential adjustment.

[0020] Third, the third-order cumulant can effectively eliminate the influence of Gaussian noise (this is because the high-order cumulant of the Gaussian process is zero), extract the phase coupling characteristics of the non-Gaussian harmonic components, effectively identify the interharmonics generated by the nonlinear load and improve the signal-to-noise ratio. At the same time, the enhanced spectrum is generated by the principal component projection, which can map the two-dimensional bispectrum to a one-dimensional spectrum, highlighting the sum frequency components of the nonlinear harmonics to improve accuracy.

[0021] Fourth, the mean of the time-frequency amplitude in the harmonic-free interval is calculated to establish a frequency-related noise floor to improve the sensitivity of high-frequency harmonic detection and effectively reduce the false positive rate compared to the fixed threshold method.

[0022] Fifth, describing the transient characteristics of harmonics through an exponential decay model can effectively avoid the fitting deviation of the traditional steady-state model for impact harmonics. At the same time, introducing frequency domain prior constraints can suppress parameter deviations caused by noise.

[0023] In general, the method of the present invention realizes a full-process closed-loop optimization from signal preprocessing to parameter optimization through a cascade architecture of adaptive time-frequency analysis (step S2), high-order statistical feature extraction (step S3), dynamic threshold detection (step S4) and prior constraint fitting (step S5). It is particularly suitable for the precise detection of heavy high-frequency, nonlinear and time-varying harmonics in vehicle-grid systems, and can provide a more reliable diagnostic basis for power quality management.

[0024] 2. Other beneficial effects or advantages of the present invention will be described in detail in the specific implementation manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] 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 description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative labor.

[0026] in: Figure 1 It is a schematic flow chart of the steps of a harmonic detection method for a vehicle grid electrical system provided by an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings 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.

[0028] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. 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.

[0029] In addition, the terms "including" and "having" and any variations thereof mentioned in the description of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes other steps or units that are not listed, or optionally includes other steps or units that are inherent to these processes, methods, products or devices. It should also be noted that in the embodiments of the present invention, words such as "exemplary" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or design. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0030] The technical solution of the present invention is described in detail below with reference to the accompanying drawings.

[0031] Example 1 like Figure 1 As shown, according to the first aspect of the present invention, this embodiment provides a harmonic detection method for a vehicle network electrical system, comprising the following steps: Step S1: obtaining a current signal of a busbar of a vehicle network and performing band-limiting processing on the current signal; Step S2: construct a frequency-adaptive variable-width S-transform kernel function; Step S3: extracting harmonic phase coupling features based on bispectral analysis of third-order cumulants, and generating enhanced spectrum through principal component projection; Step S4: Generate a dynamic detection threshold using a background noise model in a harmonic-free interval; Step S5: using an exponential decay model with frequency domain prior constraints to perform harmonic parameter fitting.

[0032] Through the above technical solution, firstly, by performing band-limiting processing on the current signal, not only can the Nyquist sampling theorem be effectively ensured to hold, but also high-frequency noise aliasing can be effectively suppressed, that is, the interference of high-frequency noise on harmonic detection can be suppressed.

[0033] Second, by constructing a frequency-adaptive variable-width S-transform kernel function, it can be well adapted to signals in different frequency bands. For low-frequency signals, the variable-width S-transform kernel function can effectively ensure its high frequency resolution. For high-frequency signals, the frequency domain leakage caused by time window widening can be suppressed through exponential adjustment.

[0034] Third, the third-order cumulant can effectively eliminate the influence of Gaussian noise (this is because the high-order cumulant of the Gaussian process is zero), extract the phase coupling characteristics of the non-Gaussian harmonic components, effectively identify the interharmonics generated by the nonlinear load and improve the signal-to-noise ratio. At the same time, the enhanced spectrum is generated by the principal component projection, which can map the two-dimensional bispectrum to a one-dimensional spectrum, highlighting the sum frequency components of the nonlinear harmonics to improve accuracy.

[0035] Fourth, the mean of the time-frequency amplitude in the harmonic-free interval is calculated to establish a frequency-related noise floor to improve the sensitivity of high-frequency harmonic detection and effectively reduce the false positive rate compared to the fixed threshold method.

[0036] Fifth, describing the transient characteristics of harmonics through an exponential decay model can effectively avoid the fitting deviation of the traditional steady-state model for impact harmonics. At the same time, introducing frequency domain prior constraints can suppress parameter deviations caused by noise.

[0037] In general, the method of the present invention realizes a full-process closed-loop optimization from signal preprocessing to parameter optimization through a cascade architecture of adaptive time-frequency analysis (step S2), high-order statistical feature extraction (step S3), dynamic threshold detection (step S4) and prior constraint fitting (step S5). It is particularly suitable for the precise detection of heavy high-frequency, nonlinear and time-varying harmonics in vehicle-grid systems, and can provide a more reliable diagnostic basis for power quality management.

[0038] In one embodiment of the present invention, step S1 specifically includes: Step S1-1: using a broadband current sensor to collect the current signal of the vehicle network bus; Step S1-2: Band-limiting processing through anti-aliasing filter: In the formula, is the complex frequency domain transfer function of the filter (used to describe the amplitude-frequency and phase-frequency response to the input signal), is a complex frequency variable, is the cut-off angular frequency (in , corresponding to the boundary of the filter passband and stopband, for example, 10kHz corresponds to ), is the filter order (determines the steepness of the transition band, the higher the order, the faster the stopband attenuation).

[0039] It can be understood that this implementation is aimed at the signal acquisition and preprocessing environment for harmonic detection of the vehicle network electrical system. By combining a broadband current sensor with a high-order anti-aliasing filter, effective retention of high-frequency harmonics and noise suppression are achieved.

[0040] Specifically, first, the cutoff frequency is The high-order filter has a smaller amplitude-frequency response flatness error in the passband, which can ensure 0- The effective harmonic components can be retained without distortion. At the same time, the stop band (greater than ) decays faster (in , For example, its stopband attenuation rate can reach -160dB / dec), which can effectively suppress high-frequency noise aliasing to avoid false harmonic components interfering with subsequent analysis and improve the accuracy and reliability of the analysis.

[0041] Second, the filter transfer function of the present invention ( ) Through high-order pole configuration, the ringing effect is reduced in the time domain, a steep transition band is achieved in the frequency domain, and the risk of high-frequency noise being folded into the low-frequency band during the sampling process is reduced.

[0042] Third, broadband current sensors (e.g., Hall effect sensors) have a wider bandwidth coverage (e.g., 0kHz-50kHz), which can meet the needs of high-frequency harmonic acquisition, and have a large signal-to-noise ratio (e.g., SNR can be greater than or equal to 80dB), which can avoid signal saturation or quantization errors.

[0043] It is understandable that the transfer function The pole symmetric distribution characteristic can ensure the maximum flat response in the passband. For example, in an exemplary embodiment, , , its amplitude-frequency response , near the cutoff frequency (e.g., ) will drop rapidly, which can effectively filter out high-frequency interference (e.g., IGBT switching noise) generated by switching devices. At the same time, the present invention is designed in collaboration with the digital sampling rate through pre-analog filtering (anti-aliasing filter) to meet the Nyquist sampling theorem. For example, if the target detection highest harmonic frequency is 10kHz, the sampling rate needs to be ≥20kHz (the actual selection can be above 50kHz), combined with the filter stopband attenuation characteristics, the aliasing component is suppressed below the system noise floor.

[0044] In one embodiment of the present invention, the cut-off angular frequency can be set to ; The filter order can be set to .

[0045] In this embodiment, the cutoff frequency is set to , in order to cover the heavy scene harmonic range and high frequency resonance components of the vehicle network system. The reason why the filter order is set to , is based on a compromise between hardware realizability (filter complexity) and stopband performance to avoid the accumulation of phase nonlinear distortion caused by high-order filters.

[0046] In general, this implementation provides a high-fidelity input signal for subsequent time-frequency analysis (such as the adaptive S-transform described below) through hardware-level signal preprocessing optimization. For example, after filtering out high-frequency noise, the window width adaptive adjustment of the S-transform kernel function in the low frequency band (f≤1kHz) can more accurately capture the interharmonic components while avoiding the waste of computational resources at invalid frequency points in the high frequency band (f≥10kHz).

[0047] In one embodiment of the present invention, step S2 may specifically include: The frequency-adaptive variable-window-width S-transform kernel function is constructed according to the following formula: In the formula, is the time-frequency matrix (representing the signal in time and frequency The energy density at For the moment The instantaneous value of the current signal at time is the frequency adaptive window width adjustment function, , is the window width ratio coefficient, is a very small frequency offset (can be set to a small value, for example, to avoid The time denominator is zero, and it does not affect the high-frequency band window width adjustment). is the Fourier basis function (to realize signal frequency decomposition).

[0048] In this implementation, the time-frequency analysis link of the harmonic detection of the vehicle network electrical system is designed through the frequency-adaptive variable window width S-transform kernel function to solve the problem of insufficient resolution in the high frequency band in the existing related technology.

[0049] Specifically, first, the window width adjustment function is used , so that the high frequency band (for example, ) window width As the frequency increases, the time domain window width decreases. , thereby improving the time-frequency energy focusing accuracy of high-frequency harmonics and optimizing the time-frequency resolution of high-frequency harmonics.

[0050] Second, through the exponential Gaussian window function ( ) dynamically adjusts the time-frequency energy distribution to suppress the spectrum leakage caused by the fixed window width in the traditional S transform in the high frequency band (for example, the diffusion of IGBT switching noise), thereby reducing the harmonic amplitude estimation error and enhancing the ability to resist spectrum leakage.

[0051] Third, the frequency-adaptive variable-width S-transform kernel function of the present invention can simultaneously capture low-frequency fundamental waves (e.g., 50 / 60 Hz) and high-frequency resonant components (e.g., 2-10 kHz PWM harmonics) to meet the harmonic detection requirements in the vehicle network system.

[0052] Among them, it can be understood that in the window width adjustment function For the low frequency band, , in this case the window width is larger (approximately equal to ), capable of high frequency resolution of low frequency harmonics (approximately ), for high frequency band, , the window width is inversely proportional to the frequency (approximately equal to ), which enables narrow time window positioning of high-frequency harmonics (for example, window width ≈ 1.6 at 10kHz).

[0053] In addition, in the frequency-adaptive variable-width S-transform kernel function, the Gaussian window and the Fourier basis ( ) work together to make the time-frequency matrix The amplitude Represents the signal at time ,frequency In this way, adaptive window width adjustment can make the energy distribution more concentrated on the real harmonic components, thereby reducing noise interference.

[0054] In general, this implementation achieves high-precision time-frequency analysis of broadband harmonics through a frequency-adaptive variable-window width S-transform kernel function, overcomes the defect of insufficient resolution of traditional methods in the high-frequency band, and provides core algorithm support for the detection of high-frequency switching noise and nonlinear harmonics in the vehicle network system.

[0055] In one embodiment of the present invention, in step S2, it also includes: When detected When the frequency adaptive window width adjustment function Adjusted to: in, is the frequency judgment threshold.

[0056] In this way, first, the time resolution of high-frequency harmonics can be improved. Specifically, when the frequency exceeds the threshold By changing the window width adjustment function from Adjust to In this way, the window width decays faster with frequency, making the high-frequency window narrower, which can significantly improve the time resolution, more accurately capture the transient changes of high-frequency harmonics (for example, fast harmonics caused by switching of power electronic devices), and reduce time ambiguity.

[0057] Second, suppress high-frequency noise interference. Specifically, the high-frequency region usually has higher noise energy, and a narrower window can reduce the signal capture time and reduce the integral effect of noise in time-frequency analysis, thereby improving the signal-to-noise ratio and enhancing the ability to identify real high-frequency harmonic components.

[0058] Third, adaptive detection requirements for different frequency bands. Specifically, low-frequency harmonics (e.g., fundamental waves, integer harmonics) in vehicle-to-grid systems are usually more stable and require wider windows to improve frequency resolution, while high-frequency harmonics (e.g., wide-spectrum switching noise) change rapidly and require higher time resolution. By dynamically adjusting the window width attenuation index, differentiated optimization of high and low frequency bands can be effectively achieved to balance the time-frequency analysis performance.

[0059] In one embodiment of the present invention, step S3 specifically includes: Step S3-1: Calculate the third-order cumulant: Step S3-2: Construct a bispectral feature matrix: Step S3-3: Implement principal component projection: In the formula, is the third-order cumulant of the signal (used to characterize the third-order statistical characteristics of the signal), is the mathematical expectation, and is the time delay, For the moment The instantaneous value of the current signal at time is the current signal at time point Shift right based on The observed value after sampling intervals, is the current signal at time point Shift right based on The observed value after sampling intervals, is the bispectral value (reflects the signal at frequency and Phase coupling characteristics at and are two independent frequency variables, is the projected spectrum (to highlight the sum frequency generated by the nonlinear system harmonic components).

[0060] Through this implementation, first, complex harmonics caused by nonlinear loads can be accurately detected. Specifically, the traditional Fourier transform cannot capture the phase coupling characteristics between harmonics, while the third-order cumulant of the present invention can effectively identify the sum frequency and difference frequency harmonic components generated by nonlinear systems (such as power electronic devices) by statistically analyzing high-order correlations.

[0061] Second, it can suppress Gaussian noise interference. Specifically, the third-order cumulant has a natural suppressive effect on Gaussian noise, and the third-order statistics of Gaussian noise in its mathematical expectation is zero. Bispectral analysis can filter out background noise and improve the accuracy of harmonic detection in low signal-to-noise ratio scenarios.

[0062] Third, it can enhance the identifiability of spectral features. Specifically, principal component projection is achieved by integrating the sum frequency components in the bispectral plane ( ), which converts complex phase-coupled signals into single-frequency energy spectra, significantly improving the spectral resolution. For example, the harmonic superposition caused by multiple inverters in parallel in a vehicle-to-grid system can be clearly separated.

[0063] In one embodiment of the present invention, step S4 specifically includes: Step S4-1: Establish background noise model: Step S4-2: Dynamic threshold calculation: In the formula, For frequency An estimate of the standard deviation of the background noise at , is the number of sampling points in the harmonic-free interval, is the time-frequency matrix at frequency ,time The amplitude at ∈ harmonic-free interval, For frequency The detection threshold at is the empirical coefficient, Set a value for the false alarm rate.

[0064] Through this implementation, first, the reliability of detection data in a noisy environment can be improved. Specifically, the background noise statistical model of the harmonic-free interval and dynamic threshold It can adapt to the noise intensity of different frequencies to avoid the misjudgment problem of fixed thresholds in the high-frequency noise range. For example, the broadband noise interference generated by power electronic switches in the vehicle-to-grid system can be effectively suppressed by this model.

[0065] Second, set the false alarm rate and empirical coefficients , which can achieve the optimal trade-off between false alarm probability and detection sensitivity in a statistical sense, and ensure the recognition accuracy of high-frequency harmonics (such as above 10kHz).

[0066] Third, it can adapt to time-varying noise environments. Specifically, the dynamic threshold is updated based on real-time collected harmonic-free interval noise data, and can cope with background noise fluctuations caused by load switching, locomotive overload, etc. during the operation of the vehicle-grid system. For example, when the motor noise suddenly increases during the acceleration of the EMU, it can still maintain stable detection performance.

[0067] In this implementation, it should be noted that extracting noise samples in a frequency band known to be free of harmonics (eg, near a low-frequency fundamental wave or a specific idle frequency band) can ensure that the noise estimation is not contaminated by harmonics.

[0068] In an exemplary practical application scenario, for example, in a high-speed rail traction power supply system, the overhead contact network voltage generates complex harmonic superposition due to the parallel operation of multiple locomotives, and the background noise contains random pulse interference (such as offline discharge of the catenary). This implementation method can distinguish between real harmonics and noise through dynamic thresholds. For example, when a certain frequency point The time-frequency energy , it can be judged as effective harmonic.

[0069] In general, this implementation solves the performance limitations of the traditional fixed threshold method in complex noise environments through dynamic noise modeling and statistical decision theory. Its core innovation lies in the adaptive estimation of background noise based on the harmonic-free interval, combined with the threshold generation mechanism constrained by the false alarm rate, which significantly improves the reliability and environmental adaptability of harmonic detection in the vehicle network electrical system.

[0070] In one embodiment of the present invention, in step S4, the empirical coefficient is set to , the false alarm rate setting value is set to .

[0071] Through this implementation, first, the balance between detection sensitivity and false alarm rate can be optimized. The coefficient amplifies the noise standard deviation estimate, combined with The false alarm rate control ensures that weak harmonics can still be effectively detected in low signal-to-noise ratio scenarios, while limiting the probability of misjudgment to less than 1%. For example, in the high-speed rail traction power supply system, it can avoid misjudgments caused by random discharge noise of the contact network (such as offline interference of the catenary), while accurately identifying low-amplitude high-frequency harmonics generated by power electronic devices.

[0072] Second, it can enhance the robustness and adaptability of the system. Specifically, the empirical coefficient Through experimental verification, it can cover the deviation between the actual noise distribution and the ideal Gaussian model, and avoid the threshold misalignment caused by changes in the statistical characteristics of the noise (such as non-Gaussian pulse interference). For example, when the load of the vehicle network system changes suddenly, the dynamic threshold can still stably distinguish between harmonics and noise.

[0073] Third, it can achieve dual optimization of high-frequency noise suppression and low-frequency harmonic retention. Specifically, the parameter combination and In view of the slow attenuation of noise energy in the high-frequency band, the detection threshold can be dynamically increased to suppress IGBT switching noise (for example, interference above 10kHz) while retaining effective signals in the low-frequency band (such as 50Hz fundamental and integer harmonics).

[0074] In one embodiment of the present invention, step S5 specifically includes: Step S5-1: Use the following algorithm to perform parameter fitting: Step S5-2: Optimize the constraints as follows: In the formula, For the The amplitude of the subharmonics, is the attenuation factor (used to describe the rate of change of harmonic amplitude over time), For the The subharmonic frequencies, is the initial phase, is the model order (i.e., the total number of harmonic components), is the number of sampling points, is the index value of the sampling point, , is the sampling interval, The current signal reconstructed by the algorithm is The predicted value of the sampling point, is the regularization coefficient, According to the a priori estimated value of the harmonic frequency obtained in steps S2 to S4.

[0075] Through this implementation, first, the accuracy of harmonic parameter estimation can be improved. Specifically, an exponential decay model with frequency domain prior constraints is used. It can accurately describe the time-varying characteristics of harmonic amplitude (such as the transient attenuation of harmonics caused by the switching of power electronic devices), and combine the prior frequency information provided by steps S2-S4 , which can significantly reduce the parameter estimation error. For example, in the high-speed rail traction system, this method can control the amplitude error of the fast-decaying high-frequency harmonics (such as above 10kHz) generated by the IGBT switch to within 1%.

[0076] Second, it can suppress noise and overfitting interference. Specifically, the optimization objective function Through the regularization term , forcing the fitting frequency to be consistent with the prior estimate to avoid the introduction of false harmonic components caused by noise interference. For example, in an exemplary embodiment, in a low signal-to-noise ratio (SNR<20dB) scenario, false harmonic misjudgments can be reduced by more than 50%.

[0077] Third, it can adapt to the time-varying characteristics of complex harmonics. Specifically, the exponential decay factor The introduction of can accurately characterize the harmonic amplitude attenuation characteristics caused by load mutation (for example, locomotive acceleration) or dynamic response of power electronic devices in the vehicle-grid system. For example, when the EMU accelerates, the motor harmonic amplitude decays exponentially with time, which cannot be captured by the traditional steady-state model, while the time-varying model provided by the present invention can achieve dynamic tracking.

[0078] Fourth, it can enhance the convergence and stability of the model. Frequency domain prior constraints The introduction of can limit the solution space of the optimization problem to the vicinity of the known harmonic frequency, avoid local optimal solutions, and accelerate convergence.

[0079] To facilitate understanding by relevant technical personnel, the following describes how to determine the frequency domain prior constraints through steps S2 to S4. The process is described below.

[0080] First, in step S2, the time-frequency matrix is ​​generated by variable window width S transformation , capturing signals at different times and frequency The energy distribution at . (For each frequency , and count the time percentage that exceeds the baseline noise in the time-frequency matrix. For example, if If the frequency is significantly higher than the noise baseline for more than 80% of the time period, it is marked as a candidate harmonic frequency. Ensure that the window width in the high frequency band is narrower, improve the time resolution of high frequency harmonics, and avoid frequency aliasing) In step S3, an enhanced spectrum is generated based on the third-order cumulant and the principal component projection. , highlighting the phase-coupled harmonics caused by nonlinear loads. (Calculate the bispectrum , identify the satisfaction The sum frequency harmonics of Enhances the energy of true harmonic frequencies and suppresses random noise. The frequencies of the top M in the medium energy ranking are regarded as high confidence candidates) In step S4, the dynamic detection threshold is calculated using the harmonic-free interval background noise model , filter out effective harmonics. (Statistical background noise standard deviation in the harmonic-free interval , reflecting the noise intensity of different frequencies, according to Generate frequency-dependent thresholds. If the time-frequency matrix of step S2 At multiple consecutive time points , and step S3 If the value is higher than the set threshold, is the real harmonic frequency, add it gather) Finally, only the frequencies that appear continuously in multiple time windows (such as sliding window analysis) are retained to avoid transient interference misjudgment. The time-frequency energy distribution of step S2 and the phase coupling characteristics of step S3 are combined to eliminate isolated frequency points (such as frequencies that appear only at a single time point). The exponential decay model of step S5 is input into the list as a frequency domain prior constraint.

[0081] In an exemplary embodiment, in a high-speed rail traction system, IGBT switches generate high-frequency harmonics (eg, 2 kHz), while catenary discharges cause random pulse noise.

[0082] In step S2, the time-frequency matrix shows that the energy at 2kHz persists, but there are instantaneous peaks in the frequency band above 10kHz.

[0083] In step S3, bispectral analysis reveals the presence of and phase coupling to enhance its spectral energy.

[0084] In step S4, the dynamic threshold determines that the 2kHz frequency exceeds 90% of the time. , while the 10kHz spike only exceeded the threshold at individual time points, so the latter was excluded.

[0085] The result is: It is input into step S5 to guide parameter fitting and accurately estimate the harmonic amplitude attenuation characteristics.

[0086] According to a second aspect of the present invention, a harmonic detection system of a vehicle grid electrical system is further provided, which is used to perform a harmonic detection method of a vehicle grid electrical system according to any technical solution in the first aspect of the present invention. The harmonic detection system of the vehicle grid electrical system comprises: The data acquisition and processing module is used to obtain the current signal of the vehicle network bus and perform band-limiting processing on the current signal; Adaptive time-frequency analysis module, used to construct frequency-adaptive variable-width S-transform kernel function; Bispectral feature extraction module, used for bispectral analysis based on third-order cumulants, extracting harmonic phase coupling features, and generating enhanced spectrum through principal component projection; A dynamic threshold generation module, used to generate a dynamic detection threshold using a harmonic-free interval background noise model; The parameter optimization module is used to fit the harmonic parameters using an exponential decay model with frequency domain prior constraints.

[0087] In this embodiment, the harmonic detection system of the vehicle-grid electrical system of the present invention seamlessly integrates signal acquisition, time-frequency analysis, feature extraction, dynamic threshold determination and parameter optimization processes through the series coordination of five functional modules, thereby significantly improving the accuracy of harmonic detection. It can also achieve full-process closed-loop optimization from signal preprocessing to parameter optimization, and is particularly suitable for the precise detection of heavy high-frequency, nonlinear and time-varying harmonics in the vehicle-grid system, and can provide a more reliable diagnostic basis for power quality management.

[0088] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for detecting harmonics in a vehicle network electrical system, characterized in that: The steps include: Step S1: obtaining a current signal of a busbar of a vehicle network and performing band-limiting processing on the current signal; Step S2: construct a frequency-adaptive variable-width S-transform kernel function; Step S3: extracting harmonic phase coupling features based on bispectral analysis of third-order cumulants, and generating enhanced spectrum through principal component projection; Step S4: Generate a dynamic detection threshold using a background noise model in a harmonic-free interval; Step S5: using an exponential decay model with frequency domain prior constraints to perform harmonic parameter fitting.

2. The harmonic detection method of the vehicle network electrical system according to claim 1, characterized in that: The step S1 specifically includes: Step S1-1: using a broadband current sensor to collect the current signal of the vehicle network bus; Step S1-2: Band-limiting processing through anti-aliasing filter: In the formula, is the complex frequency domain transfer function of the filter, is a complex frequency variable, is the cut-off angular frequency, is the filter order.

3. The harmonic detection method of the vehicle network electrical system according to claim 2, characterized in that: The cut-off frequency is set to ; The filter order is set to .

4. The harmonic detection method of the vehicle network electrical system according to claim 2, characterized in that: The step S2 specifically includes: The frequency-adaptive variable-window-width S-transform kernel function is constructed according to the following formula: In the formula, is the time-frequency matrix, For the moment The instantaneous value of the current signal at time is the frequency adaptive window width adjustment function, , is the window width ratio coefficient, is a very small frequency offset, is the Fourier basis function.

5. The harmonic detection method of the vehicle network electrical system according to claim 4, characterized in that: In the step S2, it also includes: When detected When the frequency adaptive window width adjustment function Adjusted to: in, is the frequency judgment threshold.

6. The harmonic detection method of the vehicle network electrical system according to claim 1, characterized in that: The step S3 specifically includes: Step S3-1: Calculate the third-order cumulant: Step S3-2: Construct a bispectral feature matrix: Step S3-3: Implement principal component projection: In the formula, is the third-order cumulant of the signal, is the mathematical expectation, and is the time delay, For the moment The instantaneous value of the current signal at time is the current signal at time point Shift right based on The observed value after sampling intervals, is the current signal at time point Shift right based on The observed value after sampling intervals, is the bispectral value, and are two independent frequency variables, is the projected spectrum.

7. The harmonic detection method of the vehicle network electrical system according to claim 1, characterized in that: The step S4 specifically includes: Step S4-1: Establish background noise model: Step S4-2: Dynamic threshold calculation: In the formula, For frequency An estimate of the standard deviation of the background noise at , is the number of sampling points in the harmonic-free interval, is the time-frequency matrix at frequency ,time The amplitude at ∈ harmonic-free interval, For frequency The detection threshold at is the empirical coefficient, Set a value for the false alarm rate.

8. The harmonic detection method of the vehicle network electrical system according to claim 7, characterized in that: In step S4, the empirical coefficient is set to , the false alarm rate setting value is set to .

9. The harmonic detection method of the vehicle network electrical system according to claim 1, characterized in that: The step S5 specifically includes: Step S5-1: Use the following algorithm to perform parameter fitting: Step S5-2: Optimize the constraints as follows: In the formula, For the The amplitude of the subharmonics, is the attenuation factor, For the The subharmonic frequencies, is the initial phase, is the model order, is the number of sampling points, is the index value of the sampling point, , is the sampling interval, The current signal reconstructed by the algorithm is The predicted value of the sampling points, is the regularization coefficient, According to the a priori estimated value of the harmonic frequency obtained in steps S2 to S4.

10. A harmonic detection system for a vehicle network electrical system, characterized in that: Used to perform the harmonic detection method of the vehicle-grid electrical system according to any one of claims 1 to 9, the harmonic detection system of the vehicle-grid electrical system comprises: The data acquisition and processing module is used to obtain the current signal of the vehicle network bus and perform band-limiting processing on the current signal; Adaptive time-frequency analysis module, used to construct frequency-adaptive variable-width S-transform kernel function; Bispectral feature extraction module, used for bispectral analysis based on third-order cumulants, extracting harmonic phase coupling features, and generating enhanced spectrum through principal component projection; A dynamic threshold generation module, used to generate a dynamic detection threshold using a harmonic-free interval background noise model; The parameter optimization module is used to fit the harmonic parameters using an exponential decay model with frequency domain prior constraints.

Images