High-voltage variable frequency speed regulation integrated machine fault diagnosis method and system

Abstract

The application relates to the technical field of data processing, and discloses a high-voltage variable-frequency speed-regulating integrated machine fault diagnosis method and system. The method comprises the following steps: injecting a voltage signal with a preset characteristic frequency into a frequency converter control loop to obtain a mixed voltage signal of the frequency converter control loop; separating a common-mode current characteristic waveform of a current signal in the mixed voltage signal under an excitation state; performing high-frequency resonance analysis on a vibration signal of a stator winding end to obtain an insulation defect characteristic spectrum of the stator winding end; performing time-frequency domain correlation analysis on the common-mode current characteristic waveform and the insulation defect characteristic spectrum to obtain an insulation defect characteristic matrix of the high-voltage variable-frequency speed-regulating integrated machine; and performing degradation degree evaluation on amplitude variation and distribution mode of a characteristic frequency component in the insulation defect characteristic matrix to obtain a winding insulation degradation level of the high-voltage variable-frequency speed-regulating integrated machine. The application can improve the accuracy of high-voltage variable-frequency speed-regulating integrated machine fault diagnosis.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and in particular to a high-voltage variable-frequency speed regulation integrated machine fault diagnosis method and system. BACKGROUND

[0002] In the field of high-voltage variable-frequency speed regulation integrated machine fault diagnosis, the existing technology often fails to accurately capture subtle feature changes in the detection of equipment insulation defects, resulting in a lag in fault recognition. The existing technology lacks an effective correlation mechanism in the signal processing and analysis link, and cannot deeply fuse multi-dimensional feature information, resulting in insufficient reliability of the diagnosis result and difficulty in meeting the efficient operation and maintenance requirements of the equipment.

[0003] At the same time, the existing technology has a single evaluation method for the degree of winding insulation deterioration, which relies mainly on the judgment of a single feature parameter and fails to comprehensively consider key factors such as the amplitude change and distribution pattern of the feature frequency component, resulting in low accuracy in determining the deterioration level and easy misjudgment or omission, which in turn affects the normal operation of the equipment and the timeliness of maintenance decisions. SUMMARY

[0004] The present application provides a high-voltage variable-frequency speed regulation integrated machine fault diagnosis method and system, which mainly aims to solve the problems raised in the background technology.

[0005] To achieve the above-mentioned purpose, the high-voltage variable-frequency speed regulation integrated machine fault diagnosis method provided by the present application comprises:

[0006] S1. Injecting a voltage signal of a preset feature frequency into a frequency converter control loop to obtain a mixed voltage signal of the frequency converter control loop;

[0007] S2. Separating out a common-mode current characteristic waveform of a current signal in the mixed voltage signal under excitation state;

[0008] S3. Performing high-frequency resonance analysis on the vibration signal of the stator winding end to obtain an insulation defect characteristic spectrum of the stator winding end;

[0009] S4. Performing time-frequency domain correlation analysis on the common-mode current characteristic waveform and the insulation defect characteristic spectrum to obtain an insulation defect characteristic matrix of the high-voltage variable-frequency speed regulation integrated machine;

[0010] S5. Evaluating the degree of deterioration of the amplitude change and distribution pattern of the feature frequency component in the insulation defect characteristic matrix to obtain the winding insulation deterioration level of the high-voltage variable-frequency speed regulation integrated machine.

[0011] In a preferred embodiment, the injecting of the voltage signal of the preset feature frequency into the frequency converter control loop to obtain the mixed voltage signal of the frequency converter control loop comprises:

[0012] modulating the fundamental voltage signal based on the preset characteristic frequency to obtain a modulation voltage signal of the frequency converter control loop;

[0013] modulating the modulation voltage signal and the working voltage signal to obtain a mixed voltage signal of the frequency converter control loop.

[0014] In a preferred embodiment, the separation of the common-mode current characteristic waveform of the current signal in the mixed voltage signal in the excited state includes:

[0015] extracting a common-mode component of the current signal based on the mixed voltage signal to obtain an original common-mode current of the mixed voltage signal;

[0016] performing frequency domain band-pass filtering processing on the original common-mode current to obtain an enhanced common-mode current of the mixed voltage signal;

[0017] performing time domain waveform reconstruction on the enhanced common-mode current to obtain the common-mode current characteristic waveform of the current signal in the mixed voltage signal.

[0018] In a preferred embodiment, the high-frequency resonance analysis of the vibration signal of the stator winding end portion to obtain the insulation defect characteristic spectrum of the stator winding end portion includes:

[0019] performing short-time Fourier transform on the vibration signal of the stator winding end portion to obtain a vibration spectrum diagram of the stator winding end portion;

[0020] performing resonance peak identification on the vibration spectrum diagram based on an insulation defect characteristic library to obtain a characteristic peak group of the stator winding end portion;

[0021] constructing the insulation defect characteristic spectrum of the stator winding end portion based on the frequency offset of the characteristic peak group.

[0022] In a preferred embodiment, the construction of the insulation defect characteristic spectrum of the stator winding end portion based on the frequency offset of the characteristic peak group includes:

[0023] performing discrete Fourier transform on the frequency offset of the characteristic peak group to obtain a complex spectrum value of the characteristic peak group, wherein the discrete Fourier transform calculation formula is as follows:

[0024]

[0025] wherein, is the complex spectrum value of the th frequency point, is the th value of the input sequence, is the th value of the input sequence, is the th value of the input sequence, is the total length of the input sequence, is the imaginary unit, is the frequency index, is the sequence index, is the base of the natural logarithm, is the ratio of the circumference of a circle to its diameter;

[0026] The complex spectrum values of the frequency points are aggregated as an original spectrum signal of the characteristic peak group;

[0027] The original spectrum signal is subjected to Butterworth low-pass filtering processing to obtain a filtered spectrum signal of the characteristic peak group;

[0028] The characteristic frequency points in the filtered spectrum signal are reorganized to obtain an ordered frequency point sequence of the filtered spectrum signal;

[0029] The insulation defect characteristic spectrum of the stator winding end is constructed based on the amplitude distribution of the ordered frequency point sequence.

[0030] In a preferred embodiment, the time-frequency domain correlation analysis of the common-mode current characteristic waveform and the insulation defect characteristic spectrum to obtain the insulation defect characteristic matrix of the high-voltage variable-frequency speed regulation all-in-one machine comprises:

[0031] The common-mode current characteristic waveform is subjected to short-time Fourier transform to obtain a time-frequency spectrogram of the common-mode current characteristic waveform;

[0032] The insulation defect characteristic spectrum is subjected to Hilbert transform to obtain an instantaneous frequency curve of the insulation defect characteristic spectrum;

[0033] The time-frequency spectrogram and the instantaneous frequency curve are subjected to frequency domain feature correlation analysis to obtain a correlation coefficient matrix of the high-voltage variable-frequency speed regulation all-in-one machine;

[0034] The insulation defect characteristic matrix of the high-voltage variable-frequency speed regulation all-in-one machine is constructed based on the correlation coefficient matrix.

[0035] In a preferred embodiment, the frequency domain feature correlation analysis of the time-frequency spectrogram and the instantaneous frequency curve to obtain the correlation coefficient matrix of the high-voltage variable-frequency speed regulation all-in-one machine comprises:

[0036] The time-frequency spectrogram is divided into sub-band time domain graphs according to a preset frequency band;

[0037] The instantaneous frequency segment features corresponding to the sub-band time domain graphs in the instantaneous frequency curve are extracted;

[0038] The sub-band time domain graphs and the instantaneous frequency segment features are cross-compared to obtain a frequency domain correlation feature set of the high-voltage variable-frequency speed regulation all-in-one machine;

[0039] Structuring arrangement is conducted on the frequency domain correlation feature set, to obtain a correlation coefficient matrix of the high-voltage variable-frequency speed regulation integrated machine.

[0040] In a preferred embodiment, the amplitude variation and distribution mode of the feature frequency component in the insulation defect feature matrix are evaluated to obtain a winding insulation degradation level of the high-voltage variable-frequency speed regulation integrated machine, including:

[0041] Frequency spectrum clustering analysis is conducted on the amplitude variation of the feature frequency component, to obtain a feature frequency cluster of the high-voltage variable-frequency speed regulation integrated machine.

[0042] The spatial distribution density of the feature frequency component in the feature frequency cluster is weighted by spatial density, to obtain a degradation weight coefficient of the feature frequency cluster.

[0043] Insulation degradation multi-parameter evaluation is conducted on the degradation weight coefficient and the amplitude variation amplitude of the feature frequency cluster, to obtain a winding insulation degradation level of the high-voltage variable-frequency speed regulation integrated machine.

[0044] In a preferred embodiment, the frequency spectrum clustering analysis is conducted on the amplitude variation of the feature frequency component, to obtain a feature frequency cluster of the high-voltage variable-frequency speed regulation integrated machine, including:

[0045] The amplitude-time variation curve of the feature frequency component is constructed based on the amplitude time series data of the feature frequency component.

[0046] The feature frequency component is clustered and analyzed based on the amplitude-time variation curve, to obtain an initial cluster of the feature frequency component.

[0047] Dynamic clustering optimization is conducted on the initial cluster, to obtain a feature frequency cluster of the high-voltage variable-frequency speed regulation integrated machine.

[0048] In order to solve the above problems, the application further provides a high-voltage variable-frequency speed regulation integrated machine fault diagnosis system, the system comprising:

[0049] A mixed voltage signal acquisition module is configured to inject a voltage signal of a preset feature frequency into a frequency converter control loop to obtain a mixed voltage signal of the frequency converter control loop.

[0050] A comprehensive common-mode current characteristic waveform module is configured to separate a common-mode current characteristic waveform of a current signal in the mixed voltage signal in an excitation state.

[0051] An insulation defect feature spectrum acquisition module is configured to perform high-frequency resonance analysis on a vibration signal of a stator winding end to obtain an insulation defect feature spectrum of the stator winding end.

[0052] An insulation defect feature matrix construction module is configured to perform time-frequency domain correlation analysis on the common-mode current feature waveform and the insulation defect feature spectrum to obtain an insulation defect feature matrix of the high-voltage variable-frequency speed regulation integrated machine.

[0053] A winding insulation degradation level evaluation module is configured to evaluate the amplitude variation and distribution pattern of the feature frequency components in the insulation defect feature matrix to obtain a winding insulation degradation level of the high-voltage variable-frequency speed regulation integrated machine.

[0054] Compared with the prior art, the present application has the following beneficial effects:

[0055] 1. The high-voltage variable-frequency speed regulation integrated machine fault diagnosis method and system provided by the present application can precisely capture the subtle changes of winding insulation by injecting a voltage signal of a preset feature frequency into the frequency converter control loop, performing time-frequency domain correlation analysis on the common-mode current feature waveform and the stator winding end insulation defect feature spectrum, constructing an insulation defect feature matrix, and evaluating the degradation level, thereby greatly improving the accuracy and pertinence of fault diagnosis and making the diagnosis process more scientific and effective.

[0056] 2. The present application can precisely determine the winding insulation degradation level by performing multi-dimensional analysis on the amplitude variation and distribution pattern of the feature frequency components, which can help to discover insulation defects at an early stage and evaluate the degradation level, thereby helping to take maintenance measures in a timely manner, reducing fault troubleshooting time, improving the reliability of equipment operation, and further improving the overall fault diagnosis efficiency, thereby providing a strong guarantee for the stable operation of the high-voltage variable-frequency speed regulation integrated machine. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 A flowchart of the high-voltage variable-frequency speed regulation integrated machine fault diagnosis method provided by an embodiment of the present application is shown in the figure.

[0058] Figure 2 A functional module diagram of the high-voltage variable-frequency speed regulation integrated machine fault diagnosis system provided by an embodiment of the present application is shown in the figure.

[0059] The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0060] It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0061] The embodiment of the present application provides a high-voltage variable frequency speed regulation integrated machine fault diagnosis method. The execution subject of the high-voltage variable frequency speed regulation integrated machine fault diagnosis method includes but is not limited to at least one of the electronic devices such as a server, a terminal and the like which can be configured to execute the method provided by the embodiment of the present application. In other words, the high-voltage variable frequency speed regulation integrated machine fault diagnosis method can be executed by software or hardware installed in a terminal device or a server device. The server includes but is not limited to a single server, a server cluster, a cloud server or a cloud server cluster and the like. The server can be a stand-alone server, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content distribution networks (CDN), and big data and artificial intelligence platforms and the like basic cloud computing services.

[0062] Referring to Figure 1 Fig. 1 is a flowchart of a high-voltage variable frequency speed regulation integrated machine fault diagnosis method provided by an embodiment of the present application. In the embodiment, the high-voltage variable frequency speed regulation integrated machine fault diagnosis method includes:

[0063] S1. Injecting a voltage signal of a preset characteristic frequency into a frequency converter control loop to obtain a mixed voltage signal of the frequency converter control loop;

[0064] In the embodiment of the present application, the injecting of the voltage signal of the preset characteristic frequency into the frequency converter control loop to obtain the mixed voltage signal of the frequency converter control loop includes:

[0065] Modulating a fundamental voltage signal based on the preset characteristic frequency to obtain a modulated voltage signal of the frequency converter control loop;

[0066] Modulating the modulated voltage signal and a working voltage signal to obtain the mixed voltage signal of the frequency converter control loop.

[0067] Specifically, the modulating of the fundamental voltage signal based on the preset characteristic frequency to obtain the modulated voltage signal of the frequency converter control loop.

[0068] Further, the amplitude of the fundamental voltage signal is adjusted according to the variation law of the preset characteristic frequency, so that the amplitude of the fundamental voltage signal changes with the change of the preset characteristic frequency, the frequency of the fundamental voltage signal remains unchanged, only the amplitude thereof is changed, and the voltage signal obtained after the amplitude modulation is the modulated voltage signal of the frequency converter control loop.

[0069] Further, the modulating of the modulated voltage signal and the working voltage signal to obtain the mixed voltage signal of the frequency converter control loop.

[0070] Further, the modulation voltage signal and the working voltage signal are superimposed in time to make them exist simultaneously in the same circuit, the waveform and frequency characteristics of the modulation voltage signal and the working voltage signal are kept unchanged in the superimposition process, the voltage values of the modulation voltage signal and the working voltage signal are added to obtain a new voltage signal, and the new voltage signal contains all characteristics of the modulation voltage signal and the working voltage signal, that is, a mixed voltage signal of the frequency converter control loop.

[0071] In summary, the fundamental voltage signal is modulated based on the preset characteristic frequency to obtain the modulation voltage signal of the frequency converter control loop, so that the injected voltage signal carries a specific frequency identifier, the feature response related to the insulation defect is enhanced, an explicit basis is provided for subsequent accurate capture of the fault signal, the identifiable feature of the fault is improved, and the diagnostic accuracy is improved.

[0072] In summary, the modulation voltage signal and the working voltage signal are modulated to obtain the mixed voltage signal of the frequency converter control loop, the characteristic signal can be introduced in the normal working state of the equipment, the operation of the equipment is not affected, and the mixed signal contains working state information and preset characteristic information at the same time, so that the fault diagnosis is more in line with the actual operation scene, the misjudgment caused by signal distortion is reduced, and the accuracy of fault diagnosis is further improved.

[0073] S2. Separate the common-mode current characteristic waveform of the current signal in the mixed voltage signal in the excitation state;

[0074] In the embodiment of the present application, the separation of the common-mode current characteristic waveform of the current signal in the mixed voltage signal in the excitation state comprises:

[0075] Based on the mixed voltage signal, the common-mode component of the current signal is extracted to obtain the original common-mode current of the mixed voltage signal;

[0076] The original common-mode current is subjected to frequency domain band-pass filtering to obtain the enhanced common-mode current of the mixed voltage signal;

[0077] The enhanced common-mode current is subjected to time domain waveform reconstruction to obtain the common-mode current characteristic waveform of the current signal in the mixed voltage signal.

[0078] Specifically, based on the mixed voltage signal, the common-mode component of the current signal is extracted to obtain the original common-mode current of the mixed voltage signal, specifically, the current signal values at each time point in the mixed voltage signal are added, and then the sum is divided by the number of channels of the signal to obtain the average current value at each time point, and the average current values are arranged in time sequence to form a current sequence, which is the original common-mode current of the mixed voltage signal.

[0079] Further, the original common-mode current is subjected to frequency domain band-pass filtering to obtain an enhanced common-mode current of the mixed voltage signal.

[0080] Further, the original common-mode current is converted to the frequency domain, signal components with frequencies within a preset range are identified, signal components with other frequencies are removed, and the processed frequency domain signal is converted back to the time domain to obtain a current signal, which is the enhanced common-mode current of the mixed voltage signal.

[0081] Further, the enhanced common-mode current is subjected to time domain waveform reconstruction to obtain a common-mode current characteristic waveform of the current signal in the mixed voltage signal.

[0082] Further, the value of each time point of the enhanced common-mode current is subjected to interpolation processing to increase the density of data points and make the waveform smoother, and the peak value, valley value, period, and other characteristics of the waveform are analyzed, and the waveform is adjusted and optimized according to these characteristics to make the characteristics of the waveform more obvious and prominent.

[0083] Further, the waveform obtained after such processing is the common-mode current characteristic waveform of the current signal in the mixed voltage signal.

[0084] In summary, the common-mode component extraction of the current signal based on the mixed voltage signal obtains the original common-mode current of the mixed voltage signal, which can accurately separate the common-mode component from the mixed signal, focus on the current characteristics related to insulation defects, lay a reliable foundation for subsequent analysis, reduce irrelevant signal interference, and improve the pertinence of diagnosis.

[0085] In summary, the frequency domain band-pass filtering of the original common-mode current obtains the enhanced common-mode current of the mixed voltage signal, which can filter out noise and non-target frequency components, strengthen the common-mode current characteristics related to faults, make the effective signal more prominent, reduce the influence of noise on the diagnosis result, and improve the accuracy of feature recognition.

[0086] In summary, the time domain waveform reconstruction of the enhanced common-mode current obtains the common-mode current characteristic waveform of the current signal in the mixed voltage signal, which can optimize the waveform form, clearly present the characteristic details of the common-mode current, facilitate subsequent correlation analysis with the insulation defect spectrum, enhance the accuracy of feature matching, and thus improve the accuracy of fault diagnosis.

[0087] S3. performing high-frequency resonance analysis on the vibration signal of the stator winding end to obtain an insulation defect characteristic spectrum of the stator winding end;

[0088] In the embodiment of the present application, the high-frequency resonance analysis on the vibration signal of the stator winding end to obtain the insulation defect characteristic spectrum of the stator winding end comprises:

[0089] A short-time Fourier transform is performed on the vibration signal at the end of the stator winding to obtain the vibration spectrum of the end of the stator winding.

[0090] Based on the insulation defect feature library, the vibration spectrum is used to identify resonance peaks, thereby obtaining the characteristic peak group at the end of the stator winding.

[0091] The characteristic spectrum of insulation defects at the ends of the stator winding is constructed based on the frequency offset of the characteristic peak group.

[0092] The construction of the characteristic spectrum of insulation defects at the stator winding ends based on the frequency offset of the characteristic peak group includes:

[0093] The frequency offset of the characteristic peak group is subjected to a Discrete Fourier Transform to obtain the complex spectrum value of the characteristic peak group. The calculation formula for the Discrete Fourier Transform is as follows:

[0094]

[0095] In the formula, For the first Complex spectral values ​​at each frequency point For the first input sequence One value, The total length of the input sequence. The imaginary unit, For frequency point index, For sequence index, is the base of the natural logarithm. Pi;

[0096] The complex spectral values ​​of the frequency points are collected to form the original spectral signal of the characteristic peak group;

[0097] The original spectrum signal is subjected to Butterworth low-pass filtering to obtain the filtered spectrum signal of the characteristic peak group;

[0098] The characteristic frequency points in the filtered spectrum signal are recombined to obtain an ordered frequency point sequence of the filtered spectrum signal;

[0099] The characteristic spectrum of insulation defects at the ends of the stator winding is constructed based on the amplitude distribution of the ordered frequency sequence.

[0100] Specifically, a short-time Fourier transform is performed on the vibration signal at the end of the stator winding to obtain the vibration spectrum of the end of the stator winding.

[0101] Further, the vibration signal of the stator winding end portion is divided into multiple continuous and equal length time periods in chronological order, and the vibration signal in each time period is subjected to Fourier transform to calculate different frequency components and corresponding amplitudes contained in the signal in the time period.

[0102] Further, the frequency components and amplitude information of each time period are arranged into a two-dimensional map, with time on the horizontal axis and frequency on the vertical axis, and the brightness or color of each point in the map representing the amplitude of the frequency component at the time point, and the two-dimensional map formed is the vibration spectrum of the stator winding end portion.

[0103] Further, based on the insulation defect feature library, resonance peak identification is performed on the vibration spectrum to obtain the characteristic peak group of the stator winding end portion, specifically, each frequency component and its amplitude in the vibration spectrum are compared with the standard insulation defect features pre-stored in the insulation defect feature library, and the peak value with an amplitude obviously higher than that of the surrounding frequency components in the vibration spectrum is found.

[0104] Further, it is judged whether the frequency and amplitude characteristics of these peak values match the standard features in the insulation defect feature library, and the peak values that match successfully are selected and grouped to form each set of peak values, which is the characteristic peak group of the stator winding end portion.

[0105] Further, based on the frequency offset of the characteristic peak group, the insulation defect feature spectrum of the stator winding end portion is constructed, specifically, for each characteristic peak group, the difference between the actual frequency of each peak value and the corresponding standard frequency in the insulation defect feature library is calculated, which is the frequency offset.

[0106] Further, the distribution of the frequency offset of all characteristic peak groups is analyzed, and the peak values with similar frequency offsets are classified into a class, and the number and amplitude of the peak values corresponding to each class of frequency offset are counted, and each class of frequency offset and the corresponding number and amplitude of the peak values are arranged into a spectrum according to the order of the frequency offset from small to large, which is the insulation defect feature spectrum of the stator winding end portion.

[0107] Specifically, discrete Fourier transform is performed on the frequency offset of the characteristic peak group to obtain the complex spectrum value of the characteristic peak group, specifically, all frequency offsets in the characteristic peak group are arranged in chronological order into a sequence, and the sequence is decomposed into a combination of sine and cosine waves of different frequencies, and the amplitude and phase of the sine and cosine waves corresponding to each frequency are represented by a complex number, and the complex numbers correspond to the spectrum values of each frequency point, and these values are the complex spectrum values of the characteristic peak group.

[0108] Further, the complex spectrum values of the frequency points are collected as the original spectrum signal of the characteristic peak group, specifically, the complex spectrum values corresponding to each frequency point are arranged in order from low to high, so that these values form a continuous signal sequence containing all the frequency point information, which fully reflects the spectral characteristics of the characteristic peak group at each frequency, and this sequence is the original spectrum signal of the characteristic peak group.

[0109] Further, the original spectrum signal is subjected to Butterworth low-pass filtering to obtain the filtered spectrum signal of the characteristic peak group, specifically, a fixed cutoff frequency is set, and the original spectrum signal is processed so that signal components with a frequency lower than the cutoff frequency pass through, and signal components with a frequency higher than the cutoff frequency are blocked. After such processing, the high-frequency noise components in the original spectrum signal are removed, and the remaining signal is the filtered spectrum signal of the characteristic peak group.

[0110] Further, the characteristic frequency points in the filtered spectrum signal are reorganized to obtain an ordered frequency point sequence of the filtered spectrum signal.

[0111] Further, the characteristic frequency points with an amplitude significantly higher than that of the surrounding frequency points are selected from the filtered spectrum signal, the frequency value and the corresponding amplitude of each characteristic frequency point are recorded, and these characteristic frequency points are arranged in order from low to high according to the frequency value to form a list arranged in order according to the frequency, and this list is the ordered frequency point sequence of the filtered spectrum signal.

[0112] Further, the insulation defect characteristic spectrum of the stator winding end is constructed based on the amplitude distribution of the ordered frequency point sequence.

[0113] Further, the frequency value in the ordered frequency point sequence is taken as the horizontal axis, the amplitude corresponding to each frequency value is taken as the vertical axis, and the position of each characteristic frequency point is marked in the coordinate system one by one, and these points are connected in order by lines to form a continuous curve, and the ups and downs of the curve reflect the amplitude distribution at different frequencies, and the graph constituted by the curve is the insulation defect characteristic spectrum of the stator winding end.

[0114] Specifically, is the result obtained by combining and calculating each value in the input sequence according to a specific rule, the combination rule is to multiply each input value by a complex factor related to the frequency index and the sequence index , and then add all the products to obtain the sum, which is the complex spectrum value of the corresponding frequency point . is the th value in the original input sequence, and the sequence is obtained by arranging the frequency offsets of the characteristic peak group in time order. is the input sequence The total number of values contained in the input sequence is the symbol used in mathematics to represent the imaginary unit, used to construct complex numbers. is an index used to identify different frequency points, starting from 0 and increasing sequentially until Each value corresponds to a specific frequency point. is an index used to identify the position of each value in the input sequence, starting from 0 and increasing sequentially until . is the base of the natural logarithm, a fixed mathematical constant. is the ratio of the circumference of a circle to its diameter, also a fixed mathematical constant.

[0115] Further, the meaning of the formula is to convert the input sequence in the time domain to the frequency domain, obtaining a series of complex spectrum values Each complex spectrum value represents the amplitude and phase information of the frequency component of the original sequence at the corresponding frequency point Through this conversion, we can analyze which frequency components are contained in the original sequence and the strength of each frequency component, thus more clearly understanding the frequency characteristics of the signal.

[0116] Further, when the values in the input sequence change rapidly, the amplitude of the complex spectrum value in the high-frequency region will be relatively large, indicating that the original sequence contains more high-frequency components; when the values in the input sequence change slowly, the amplitude of the complex spectrum value in the low-frequency region will be relatively large, indicating that the original sequence mainly contains low-frequency components; if the values in the input sequence do not change with time, most of the complex spectrum values will be concentrated in the low-frequency region with large amplitude, while the amplitude in the high-frequency region will be very small.

[0117] In summary, performing short-time Fourier transform on the vibration signal of the stator winding end to obtain the vibration spectrum of the stator winding end can convert the vibration signal into a time-frequency two-dimensional map, clearly presenting the frequency components and amplitude changes at different times, accurately capturing the high-frequency resonance characteristics related to insulation defects, providing detailed signal basis for fault diagnosis, and improving the accuracy of feature recognition.

[0118] In general, the resonance peak recognition is performed on the vibration spectrum based on the insulation defect feature library to obtain the characteristic peak group of the stator winding end, the key peak value in the spectrum can be screened by means of the known defect feature, irrelevant peak value interference is excluded, the characteristic peak directly related to the insulation defect is focused, the extracted feature has clear directivity, and the influence of invalid information on diagnosis is reduced.

[0119] In general, the insulation defect feature spectrum of the stator winding end is constructed based on the frequency offset of the characteristic peak group, the subtle change of the insulation defect can be reflected through the quantitative index of frequency offset, the dispersed characteristic peak is converted into a structured spectrum, the distribution rule of the defect feature is clearly presented, the basis for diagnosis is more systematic and quantitative, and the accuracy of fault diagnosis is further improved.

[0120] In general, the discrete Fourier transform is performed on the frequency offset of the characteristic peak group to obtain the complex spectrum value of the characteristic peak group, the time domain signal of the frequency offset can be converted into the complex representation in the frequency domain, the amplitude and phase information of the frequency component are completely retained, the frequency domain feature of the characteristic peak group is accurately captured, comprehensive basic data is provided for subsequent analysis, and the accuracy of diagnosis is improved.

[0121] In general, the complex spectrum values of the frequency points are collected as the original spectrum signal of the characteristic peak group, the information of each frequency point can be integrated to form a continuous spectrum signal, the overall frequency domain distribution of the characteristic peak group is completely presented, the feature omission caused by dispersed information is avoided, the subsequent processing is based on a complete feature set, and the diagnosis deviation is reduced.

[0122] In general, the Butterworth low-pass filtering processing is performed on the original spectrum signal to obtain the filtered spectrum signal of the characteristic peak group, the high-frequency noise interference can be effectively filtered out, the low-frequency characteristic signal related to the insulation defect is retained, the spectrum signal is purer, the interference of noise on feature recognition is reduced, and the accuracy of feature extraction is improved.

[0123] In general, the characteristic frequency points in the filtered spectrum signal are reorganized to obtain an ordered frequency point sequence of the filtered spectrum signal, the characteristic frequency points can be combed according to a rule, the frequency points are arranged in order, the distribution logic of the characteristic frequency is clearly presented, the key defect feature can be quickly located, and the pertinence of diagnosis is enhanced.

[0124] In general, the insulation defect feature spectrum of the stator winding end is constructed based on the amplitude distribution of the ordered frequency point sequence, the amplitude distribution of the ordered frequency point can be converted into a structured spectrum, the correlation between the frequency and the amplitude of the insulation defect feature is intuitively reflected, a clear and quantitative feature basis is provided for fault diagnosis, and the diagnosis accuracy is further improved.

[0125] S4. performing time-frequency domain correlation analysis on the common-mode current characteristic waveform and the insulation defect characteristic spectrum to obtain an insulation defect characteristic matrix of the high-voltage variable-frequency speed-regulating integrated machine;

[0126] In the embodiment of the present application, the time-frequency domain correlation analysis on the common-mode current characteristic waveform and the insulation defect characteristic spectrum to obtain an insulation defect characteristic matrix of the high-voltage variable-frequency speed-regulating integrated machine comprises:

[0127] performing short-time Fourier transform on the common-mode current characteristic waveform to obtain a time-frequency spectrogram of the common-mode current characteristic waveform;

[0128] performing Hilbert transform on the insulation defect characteristic spectrum to obtain an instantaneous frequency curve of the insulation defect characteristic spectrum;

[0129] performing frequency domain feature correlation analysis on the time-frequency spectrogram and the instantaneous frequency curve to obtain a correlation coefficient matrix of the high-voltage variable-frequency speed-regulating integrated machine;

[0130] constructing the insulation defect characteristic matrix of the high-voltage variable-frequency speed-regulating integrated machine based on the correlation coefficient matrix.

[0131] The frequency domain feature correlation analysis on the time-frequency spectrogram and the instantaneous frequency curve to obtain a correlation coefficient matrix of the high-voltage variable-frequency speed-regulating integrated machine comprises:

[0132] dividing the time-frequency spectrogram into sub-band time domain graphs according to a preset frequency band;

[0133] extracting an instantaneous frequency segment feature in the instantaneous frequency curve corresponding to the sub-band time domain graph;

[0134] crossing and comparing the sub-band time domain graph and the instantaneous frequency segment feature to obtain a frequency domain correlation feature set of the high-voltage variable-frequency speed-regulating integrated machine;

[0135] performing structured arrangement on the frequency domain correlation feature set to obtain a correlation coefficient matrix of the high-voltage variable-frequency speed-regulating integrated machine.

[0136] Specifically, the short-time Fourier transform is performed on the common-mode current characteristic waveform to obtain a time-frequency spectrogram of the common-mode current characteristic waveform, specifically, the common-mode current characteristic waveform is divided into a plurality of time periods with equal length, the Fourier transform is performed on the waveform in each time period, the different frequency components and the corresponding amplitudes contained in the waveform in the time period are calculated, and the frequency components and the amplitude information of each time period are arranged into a two-dimensional graph, in which the horizontal axis represents time and the vertical axis represents frequency, the brightness or color of each point in the graph represents the amplitude of the frequency component at the time point, and the two-dimensional graph formed is the time-frequency spectrogram of the common-mode current characteristic waveform.

[0137] Further, the insulation defect feature spectrum is subjected to Hilbert transform to obtain an instantaneous frequency curve of the insulation defect feature spectrum. Specifically, the insulation defect feature spectrum is subjected to inverse Fourier transform to convert it from the frequency domain back to a time domain signal, the obtained time domain signal is subjected to Hilbert transform to construct an analytic signal of the signal, the instantaneous frequency of each time point is calculated through the phase change rate of the analytic signal, and the instantaneous frequencies are connected in time sequence to form a continuous curve, which is the instantaneous frequency curve of the insulation defect feature spectrum.

[0138] Further, the time-frequency spectrogram and the instantaneous frequency curve are subjected to frequency domain feature correlation analysis to obtain a correlation coefficient matrix of the high-voltage variable-frequency speed regulation integrated machine. Specifically, the main frequency components and their amplitudes of each time point are extracted from the time-frequency spectrogram, the instantaneous frequency values of each time point are extracted from the instantaneous frequency curve, the frequency components in the time-frequency spectrogram and the instantaneous frequency values in the instantaneous frequency curve at the same time point are compared, the similarity and consistency of the change trend therebetween are evaluated, and a numerical value is used to represent the correlation. The frequency components and the instantaneous frequency values of all time points are subjected to such correlation evaluation, and all the obtained correlation numerical values are arranged into a matrix, the rows and columns of the matrix correspond to different time points and frequency components respectively, and the matrix is the correlation coefficient matrix of the high-voltage variable-frequency speed regulation integrated machine.

[0139] Further, the insulation defect feature matrix of the high-voltage variable-frequency speed regulation integrated machine is constructed based on the correlation coefficient matrix. Specifically, the correlation coefficient matrix is subjected to normalization processing, so that all the numerical values in the matrix are within the range of 0 to 1, and the comparability of the numerical values is enhanced. According to the sizes of the normalized correlation coefficients, the correlation degrees of each frequency component and the insulation defect are determined. The higher the correlation coefficient is, the stronger the correlation between the frequency component and the insulation defect is. The frequency components with high correlation degrees and their corresponding correlation coefficients are arranged into a new matrix according to certain rules, and the matrix is the insulation defect feature matrix of the high-voltage variable-frequency speed regulation integrated machine.

[0140] Specifically, the time-frequency spectrogram is divided into sub-band time domain graphs according to preset frequency bands. Specifically, a plurality of preset frequency band ranges are determined, each frequency band range contains a certain frequency interval, all the frequency components and their corresponding amplitude information belonging to the same frequency band range in the time-frequency spectrogram are extracted, and are recombined in the original time sequence to form a new time domain graph containing only the frequency components in the frequency band range. The extraction and recombination operation is performed on each preset frequency band to obtain a plurality of new time domain graphs, which are the sub-band time domain graphs.

[0141] Further, the instantaneous frequency segment features corresponding to the sub-band time-domain graph in the instantaneous frequency curve are extracted, specifically, on the instantaneous frequency curve, a part with the same time range as each sub-band time-domain graph is found, the part of the curve is separately intercepted, the features of the intercepted curve segment are analyzed, including the average frequency value of the curve, the amplitude range of the frequency change, the rate of the frequency change, and the trend of the frequency change, etc., and the features obtained by the analysis are arranged into a set of feature descriptions, and the set of feature descriptions is the instantaneous frequency segment feature corresponding to the sub-band time-domain graph.

[0142] Further, the sub-band time-domain graph and the instantaneous frequency segment feature are cross-compared to obtain a frequency domain correlation feature set of the high-voltage variable frequency speed regulation integrated machine, specifically, the frequency component and the amplitude value change in each sub-band time-domain graph are compared in detail with the corresponding instantaneous frequency segment feature, whether the frequency component change in the sub-band time-domain graph at the same time point is consistent with the frequency change trend in the instantaneous frequency segment feature is observed, the similarity degree and the correlation strength between the two are evaluated, and the comparison results of all the sub-band time-domain graphs and the corresponding instantaneous frequency segment features are arranged into a set, and the set is the frequency domain correlation feature set of the high-voltage variable frequency speed regulation integrated machine.

[0143] Further, the frequency domain correlation feature set is structured and arranged to obtain a correlation coefficient matrix of the high-voltage variable frequency speed regulation integrated machine, specifically, each correlation feature in the frequency domain correlation feature set is classified and sorted according to certain rules, the position of each correlation feature in the matrix is determined, a numerical value representing the correlation coefficient is given to each correlation feature according to the similarity degree and the correlation strength of the correlation feature, the correlation coefficients corresponding to all the correlation features are filled into a matrix according to the determined positions, and a regularly arranged matrix table is formed, which is the correlation coefficient matrix of the high-voltage variable frequency speed regulation integrated machine.

[0144] In summary, the short-time Fourier transform of the common-mode current characteristic waveform is performed to obtain a time-frequency spectrogram of the common-mode current characteristic waveform, which can convert the time domain features of the common-mode current into time-frequency two-dimensional features, clearly present the frequency changes at different times, accurately capture the dynamic features related to the insulation defect, provide detailed time-frequency information for correlation analysis, and improve the accuracy of feature matching.

[0145] In summary, the Hilbert transform of the insulation defect characteristic frequency spectrum is performed to obtain an instantaneous frequency curve of the insulation defect characteristic frequency spectrum, which can extract the instantaneous frequency features of the frequency spectrum changing over time, reflect the dynamic evolution law of the insulation defect, convert the static frequency spectrum into a dynamic feature curve, enhance the correlation with the time-frequency features of the common-mode current, and provide a dynamic basis for subsequent correlation analysis.

[0146] Overall, the frequency domain feature correlation analysis of the time-frequency spectrogram and the instantaneous frequency curve obtains the correlation coefficient matrix of the high-voltage variable frequency speed regulation integrated machine, energy quantizes the correlation degree of both in the frequency domain, clearly defines the matching relationship of different frequency components, filters irrelevant correlations, focuses on the feature association directly related to insulation defects, and reduces analysis deviation.

[0147] Overall, constructing the insulation defect feature matrix of the high-voltage variable frequency speed regulation integrated machine based on the correlation coefficient matrix can integrate multi-dimensional associated features into a structured matrix, the system presents the comprehensive features of insulation defects, makes the diagnosis basis more comprehensive and clear, and effectively improves the accuracy of fault diagnosis.

[0148] When the time-frequency spectrogram is divided into sub-band time-domain graphs according to a preset frequency band, the complex time-frequency information can be refined by frequency band, the feature changes in each frequency band are focused, information confusion in overall analysis is avoided, a foundation is laid for accurate correlation analysis, the pertinence of feature extraction is improved, and the diagnostic accuracy is improved.

[0149] Overall, extracting the instantaneous frequency segment features corresponding to the sub-band time-domain graphs in the instantaneous frequency curve can accurately correspond the instantaneous frequency features to the sub-band time-domain graphs in time and frequency band, ensure that the comparison of the two is carried out in the same dimension, reduce the correlation deviation caused by feature misplacement, and enhance the accuracy of feature correlation.

[0150] Overall, cross-comparing the sub-band time-domain graphs and the instantaneous frequency segment features obtains a frequency domain associated feature set of the high-voltage variable frequency speed regulation integrated machine, which can comprehensively mine the correlation relationship between time-frequency features and instantaneous frequency features in different frequency bands, capture subtle correlation differences, enrich the dimension of associated features, and provide a more comprehensive basis for diagnosis.

[0151] Overall, structuring the frequency domain associated feature set obtains a correlation coefficient matrix of the high-voltage variable frequency speed regulation integrated machine, which can systematically and quantitatively present the dispersed associated features, clearly reflect the correlation strength between features, facilitate quick identification of key correlations, reduce analysis errors, and further improve the accuracy of fault diagnosis.

[0152] S5. The amplitude variation and distribution mode of the feature frequency component in the insulation defect feature matrix are evaluated for degradation degree, and a winding insulation degradation level of the high-voltage variable frequency speed regulation integrated machine is obtained.

[0153] In the embodiment of the present application, the degradation degree evaluation of the amplitude variation and distribution mode of the feature frequency component in the insulation defect feature matrix obtains the winding insulation degradation level of the high-voltage variable frequency speed regulation integrated machine, which comprises:

[0154] performing spectral clustering analysis on the amplitude variation of the characteristic frequency components to obtain a characteristic frequency cluster of the high-voltage variable-frequency speed regulation integrated machine;

[0155] performing spatial density weighting on the spatial distribution density of the characteristic frequency components in the characteristic frequency cluster to obtain a degradation weight coefficient of the characteristic frequency cluster;

[0156] performing insulation degradation multi-parameter evaluation on the degradation weight coefficient and the amplitude variation amplitude of the characteristic frequency cluster to obtain a winding insulation degradation level of the high-voltage variable-frequency speed regulation integrated machine.

[0157] The performing spectral clustering analysis on the amplitude variation of the characteristic frequency components to obtain a characteristic frequency cluster of the high-voltage variable-frequency speed regulation integrated machine comprises:

[0158] constructing an amplitude-time variation curve of the characteristic frequency components based on the amplitude time series data of the characteristic frequency components;

[0159] performing clustering analysis on the characteristic frequency components based on the amplitude-time variation curve to obtain an initial cluster of the characteristic frequency components;

[0160] performing dynamic clustering optimization on the initial cluster to obtain a characteristic frequency cluster of the high-voltage variable-frequency speed regulation integrated machine.

[0161] Specifically, the performing spectral clustering analysis on the amplitude variation of the characteristic frequency components to obtain a characteristic frequency cluster of the high-voltage variable-frequency speed regulation integrated machine specifically comprises extracting amplitude variation data of all characteristic frequency components from an insulation defect characteristic matrix, recording amplitude variation trends of each frequency component at different times, grouping frequency components with the same or similar variation trends, comparing amplitude variation rules of frequency components in each group to ensure that variation modes of components in the group are consistent and variation modes of components between groups are obviously different, and finally forming each group as a characteristic frequency cluster of the high-voltage variable-frequency speed regulation integrated machine.

[0162] Further, the performing spatial density weighting on the spatial distribution density of the characteristic frequency components in the characteristic frequency cluster to obtain a degradation weight coefficient of the characteristic frequency cluster specifically comprises: statistically analyzing spatial distribution positions of all characteristic frequency components in each characteristic frequency cluster in the insulation structure, calculating the number of the cluster characteristic frequency components contained in a unit spatial range, taking the number as the spatial distribution density, setting a weighting rule according to the size of the spatial distribution density, and giving each characteristic frequency cluster a corresponding numerical value according to the rule, the numerical value being the degradation weight coefficient of the characteristic frequency cluster.

[0163] Further, insulation deterioration multi-parameter evaluation is performed on the amplitude variation range of the amplitude of the deterioration weight coefficient and the characteristic frequency cluster to obtain the winding insulation deterioration grade of the high-voltage variable-frequency speed regulation all-in-one machine. Specifically, the deterioration weight coefficient of each characteristic frequency cluster is multiplied by the amplitude variation range of the cluster to obtain an insulation deterioration evaluation value of each cluster. The evaluation values of all characteristic frequency clusters are summarized to obtain a total evaluation result. The total evaluation result is compared with a preset winding insulation deterioration grade division standard. The interval grade corresponding to the interval in which the total evaluation result falls is the winding insulation deterioration grade of the high-voltage variable-frequency speed regulation all-in-one machine.

[0164] Specifically, the amplitude data of each characteristic frequency component at different time points is obtained to form the amplitude time sequence data of the characteristic frequency component. With time as the horizontal axis and amplitude as the vertical axis, the amplitude data corresponding to each time point is marked in the coordinate system, and then the marked points are sequentially connected in time order to form a continuous line, which is the amplitude-time variation curve of the characteristic frequency component.

[0165] Further, the amplitude-time variation curves of all characteristic frequency components are compared to observe the overall trend, fluctuation range and change rhythm of the curves. The characteristic frequency components with similar curve shapes and consistent change trends are grouped into the same group, so that the curves of the components in the group remain highly consistent in change mode, and the curves of the components in different groups are significantly different in change mode. Each group obtained after such grouping is an initial cluster of characteristic frequency components.

[0166] Further, the characteristic frequency components in each initial cluster are checked one by one, and their amplitude-time variation curves are recompared. If it is found that the curve of a component is significantly different from the curves of other components in the cluster, the component is removed from the initial cluster and is re-assigned to a more matched initial cluster. Meanwhile, different initial clusters with extremely similar curve characteristics are merged, so that the curve characteristics of the components in each adjusted cluster are completely consistent. Finally, the stable grouping obtained is the characteristic frequency cluster of the high-voltage variable-frequency speed regulation all-in-one machine.

[0167] In summary, the amplitude variation of the characteristic frequency component is subjected to spectral clustering analysis to obtain the characteristic frequency cluster of the high-voltage variable-frequency speed regulation all-in-one machine. The frequency components with similar change trends are classified, the overall change law of the same type of characteristics is focused, single characteristic interference is avoided, the analysis is more targeted, and the recognition accuracy of the insulation deterioration characteristics is improved.

[0168] In summary, the spatial distribution density of the characteristic frequency components in the characteristic frequency cluster is subjected to spatial density weighting to obtain the deterioration weight coefficient of the characteristic frequency cluster. Different weights can be given according to the distribution density of the characteristics in the insulation structure, the deterioration characteristics of the key region are highlighted, the rationality of the evaluation is enhanced, and the influence of the secondary characteristics on the result is reduced.

[0169] In summary, the insulation deterioration multi-parameter evaluation is performed on the amplitude variation range of the deterioration weight coefficient and the amplitude of the characteristic frequency cluster to obtain the winding insulation deterioration grade of the high-voltage variable-frequency speed regulation integrated machine, which can comprehensively consider the distribution importance and variation degree of the characteristics, realize multi-dimensional quantitative evaluation, avoid the limitation of single parameter judgment, and greatly improve the accuracy of winding insulation deterioration grade determination.

[0170] In summary, the amplitude-time variation curve of the characteristic frequency component is constructed based on the amplitude time series data of the characteristic frequency component, which can intuitively present the amplitude fluctuation law of each frequency component over time, clearly reflect the dynamic characteristics related to insulation deterioration, provide visual basis for subsequent clustering analysis, help to accurately identify the characteristic change mode, and improve the reliability of the diagnostic basic data.

[0171] In summary, the initial cluster of the characteristic frequency component is obtained by clustering analysis of the characteristic frequency component based on the amplitude-time variation curve, which can group the frequency components with similar change trends into a group, focus on the overall change of the same characteristics, reduce the interference of scattered characteristics on analysis, make the classification of insulation deterioration related characteristics more targeted, and lay a grouping foundation for accurate evaluation.

[0172] In summary, the characteristic frequency cluster of the high-voltage variable-frequency speed regulation integrated machine is obtained by dynamic clustering optimization of the initial cluster, which can eliminate the deviation in the initial grouping by adjustment and optimization, ensure that the characteristic change mode in the cluster is highly consistent and the difference between clusters is significant, make the characteristic frequency cluster more accurately reflect the actual characteristics of insulation deterioration, and further improve the accuracy of fault diagnosis.

[0173] As shown in Figure 2 FIG. 1 is a functional module diagram of a high-voltage variable-frequency speed regulation integrated machine fault diagnosis system according to an embodiment of the present application.

[0174] The high-voltage variable-frequency speed regulation integrated machine fault diagnosis system 100 can be installed in an electronic device. According to the functions implemented, the high-voltage variable-frequency speed regulation integrated machine fault diagnosis system 100 can include a hybrid voltage signal acquisition module 101, a comprehensive common-mode current characteristic waveform module 102, an insulation defect characteristic spectrum acquisition module 103, an insulation defect characteristic matrix construction module 104, and a winding insulation deterioration grade evaluation module 105. The modules of the present application can also be referred to as units, which refer to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, and are stored in the memory of the electronic device.

[0175] In this embodiment, the functions of each module / unit are as follows:

[0176] The mixed voltage signal acquisition module 101 is configured to inject a voltage signal with a preset characteristic frequency into a frequency converter control loop to obtain a mixed voltage signal of the frequency converter control loop.

[0177] The comprehensive common-mode current characteristic waveform module 102 is configured to separate a common-mode current characteristic waveform of a current signal in the mixed voltage signal in an excitation state.

[0178] The insulation defect characteristic frequency spectrum acquisition module 103 is configured to perform high-frequency resonance analysis on a vibration signal of a stator winding end to obtain an insulation defect characteristic frequency spectrum of the stator winding end.

[0179] The insulation defect characteristic matrix construction module is configured to perform time-frequency domain correlation analysis 104 on the common-mode current characteristic waveform and the insulation defect characteristic frequency spectrum to obtain an insulation defect characteristic matrix of the high-voltage variable-frequency speed regulation integrated machine.

[0180] The winding insulation degradation level evaluation module 105 is configured to evaluate a degradation degree of a variation in amplitude and a distribution mode of a characteristic frequency component in the insulation defect characteristic matrix to obtain a winding insulation degradation level of the high-voltage variable-frequency speed regulation integrated machine.

[0181] In several embodiments provided in the present application, it should be understood that the disclosed method and system can be implemented in other ways. For example, the system embodiments described above are merely illustrative, for example, the division of the modules is only a logical function division, and another division mode can be used in actual implementation.

[0182] The modules described as separate components can or can not be physically separated, and the components shown as modules can or can not be physical units, that is, they can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment according to actual needs.

[0183] In addition, each functional module in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of hardware plus software functional modules.

[0184] It is obvious for those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and the present application can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application.

[0185] The embodiments of the present application can acquire and process related data based on artificial intelligence technology. The artificial intelligence is a theory, method, technology and application system for simulating, extending and expanding human intelligence by using a digital computer or a machine controlled by a digital computer, perceiving an environment, acquiring knowledge and using the knowledge to obtain optimal results.

[0186] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the present application.

Claims

1. A fault diagnosis method for a high-voltage variable frequency speed control integrated machine, characterized in that, The method includes: S1. Inject a voltage signal with a preset characteristic frequency into the inverter control circuit to obtain a mixed voltage signal of the inverter control circuit; S2. Separate the common-mode current characteristic waveform of the current signal in the mixed voltage signal under the excitation state, including: Based on the mixed voltage signal, the common-mode component of the current signal is extracted to obtain the original common-mode current of the mixed voltage signal; The original common-mode current is subjected to frequency domain bandpass filtering to obtain the enhanced common-mode current of the mixed voltage signal; The enhanced common-mode current is reconstructed in the time domain to obtain the common-mode current characteristic waveform of the current signal in the hybrid voltage signal; S3. Perform high-frequency resonance analysis on the vibration signal at the end of the stator winding to obtain the characteristic spectrum of insulation defects at the end of the stator winding, including: A short-time Fourier transform is performed on the vibration signal at the end of the stator winding to obtain the vibration spectrum of the end of the stator winding. Based on the insulation defect feature library, the vibration spectrum is used to identify resonance peaks, thereby obtaining the characteristic peak group at the end of the stator winding. The characteristic spectrum of insulation defects at the stator winding ends is constructed based on the frequency offset of the characteristic peak group, including: The frequency offset of the characteristic peak group is subjected to a Discrete Fourier Transform to obtain the complex spectrum value of the characteristic peak group. The calculation formula for the Discrete Fourier Transform is as follows: ； In the formula, For the first Complex spectral values ​​at each frequency point For the first input sequence One value, The total length of the input sequence. The imaginary unit, For frequency point index, For sequence index, is the base of the natural logarithm. Pi; The complex spectral values ​​of the frequency points are collected to form the original spectral signal of the characteristic peak group; The original spectrum signal is subjected to Butterworth low-pass filtering to obtain the filtered spectrum signal of the characteristic peak group; The characteristic frequency points in the filtered spectrum signal are recombined to obtain an ordered frequency point sequence of the filtered spectrum signal; The characteristic spectrum of insulation defects at the ends of the stator winding is constructed based on the amplitude distribution of the ordered frequency point sequence. S4. Perform time-frequency domain correlation analysis on the common-mode current characteristic waveform and the insulation defect characteristic spectrum to obtain the insulation defect characteristic matrix of the high-voltage variable frequency speed control integrated machine, including: Perform a short-time Fourier transform on the common-mode current characteristic waveform to obtain the time-frequency spectrum of the common-mode current characteristic waveform; The Hilbert transform is applied to the characteristic spectrum of the insulation defect to obtain the instantaneous frequency curve of the characteristic spectrum of the insulation defect; Frequency domain correlation analysis was performed on the time-frequency spectrum and the instantaneous frequency curve to obtain the correlation coefficient matrix of the high-voltage variable frequency speed control integrated machine; Based on the correlation coefficient matrix, an insulation defect feature matrix of the high-voltage variable frequency speed control integrated machine is constructed; S5. Evaluate the degree of degradation of the amplitude variation and distribution pattern of the characteristic frequency components in the insulation defect feature matrix to obtain the winding insulation degradation level of the high-voltage variable frequency speed control integrated machine.

2. The fault diagnosis method for a high-voltage variable frequency speed control integrated machine as described in claim 1, characterized in that, The process of injecting a voltage signal with a preset characteristic frequency into the inverter control circuit to obtain a mixed voltage signal for the inverter control circuit includes: The fundamental voltage signal is modulated based on the preset characteristic frequency to obtain the modulated voltage signal of the inverter control circuit; The modulation voltage signal and the working voltage signal are modulated to obtain the mixed voltage signal of the inverter control circuit.

3. The fault diagnosis method for a high-voltage variable frequency speed control integrated machine as described in claim 1, characterized in that, The frequency domain correlation analysis of the time-frequency spectrum and the instantaneous frequency curve yields the correlation coefficient matrix of the high-voltage variable frequency speed control integrated machine, including: The time-frequency spectrum is divided into sub-frequency band time-domain graphs according to a preset frequency band. Extract the instantaneous frequency segmentation features from the instantaneous frequency curve that correspond to the time domain plot of the sub-frequency band; By cross-comparing the sub-frequency band time domain diagram with the instantaneous frequency segmentation features, the frequency domain correlation feature set of the high-voltage variable frequency speed control integrated machine is obtained; The correlation coefficient matrix of the high-voltage variable frequency speed control integrated machine is obtained by structuring the set of frequency domain correlation features.

4. The fault diagnosis method for a high-voltage variable frequency speed control integrated machine as described in claim 1, characterized in that, The degradation level of the winding insulation of the high-voltage variable frequency speed control unit is obtained by evaluating the amplitude variation and distribution pattern of the characteristic frequency components in the insulation defect feature matrix, including: Spectral clustering analysis is performed on the amplitude changes of the characteristic frequency components to obtain the characteristic frequency clusters of the high-voltage variable frequency speed control integrated machine; Spatial density weighting is applied to the spatial distribution density of the characteristic frequency components in the characteristic frequency cluster to obtain the degradation weight coefficient of the characteristic frequency cluster. The insulation degradation level of the high-voltage variable frequency speed control integrated machine is obtained by performing a multi-parameter evaluation on the degradation weight coefficient and the amplitude change of the characteristic frequency cluster.

5. The fault diagnosis method for a high-voltage variable frequency speed control integrated machine as described in claim 4, characterized in that, The spectral clustering analysis of the amplitude changes of the characteristic frequency components yields the characteristic frequency clusters of the high-voltage variable frequency speed control integrated machine, including: Construct the amplitude-time variation curve of the characteristic frequency component based on the amplitude time series data of the characteristic frequency component; Cluster analysis is performed on the characteristic frequency components based on the amplitude-time variation curve to obtain the initial clusters of the characteristic frequency components; Dynamic clustering optimization is performed on the initial cluster to obtain the characteristic frequency cluster of the high-voltage variable frequency speed control integrated machine.

6. A fault diagnosis system for a high-voltage variable frequency speed control integrated machine, used to implement the fault diagnosis method for a high-voltage variable frequency speed control integrated machine as described in any one of claims 1-5, characterized in that, The system includes: A hybrid voltage signal acquisition module is used to inject a voltage signal with a preset characteristic frequency into the inverter control circuit to obtain a hybrid voltage signal of the inverter control circuit. The integrated common-mode current characteristic waveform module is used to separate the common-mode current characteristic waveform of the current signal in the mixed voltage signal under excitation state, including: Based on the mixed voltage signal, the common-mode component of the current signal is extracted to obtain the original common-mode current of the mixed voltage signal; The original common-mode current is subjected to frequency domain bandpass filtering to obtain the enhanced common-mode current of the mixed voltage signal; The enhanced common-mode current is reconstructed in the time domain to obtain the common-mode current characteristic waveform of the current signal in the hybrid voltage signal; An insulation defect characteristic spectrum acquisition module is used to perform high-frequency resonance analysis on the vibration signal at the end of the stator winding to obtain the characteristic spectrum of insulation defects at the end of the stator winding, including: A short-time Fourier transform is performed on the vibration signal at the end of the stator winding to obtain the vibration spectrum of the end of the stator winding. Based on the insulation defect feature library, the vibration spectrum is used to identify resonance peaks, thereby obtaining the characteristic peak group at the end of the stator winding. The characteristic spectrum of insulation defects at the stator winding ends is constructed based on the frequency offset of the characteristic peak group, including: The frequency offset of the characteristic peak group is subjected to a Discrete Fourier Transform to obtain the complex spectrum value of the characteristic peak group. The calculation formula for the Discrete Fourier Transform is as follows: ； In the formula, For the first Complex spectral values ​​at each frequency point For the first input sequence One value, The total length of the input sequence. The imaginary unit, For frequency point index, For sequence index, is the base of the natural logarithm. Pi; The complex spectral values ​​of the frequency points are collected to form the original spectral signal of the characteristic peak group; The original spectrum signal is subjected to Butterworth low-pass filtering to obtain the filtered spectrum signal of the characteristic peak group; The characteristic frequency points in the filtered spectrum signal are recombined to obtain an ordered frequency point sequence of the filtered spectrum signal; The characteristic spectrum of insulation defects at the ends of the stator winding is constructed based on the amplitude distribution of the ordered frequency point sequence. An insulation defect feature matrix construction module is used to perform time-frequency domain correlation analysis between the common-mode current feature waveform and the insulation defect feature spectrum to obtain the insulation defect feature matrix of the high-voltage variable frequency speed control integrated machine, including: Perform a short-time Fourier transform on the common-mode current characteristic waveform to obtain the time-frequency spectrum of the common-mode current characteristic waveform; The Hilbert transform is applied to the characteristic spectrum of the insulation defect to obtain the instantaneous frequency curve of the characteristic spectrum of the insulation defect; Frequency domain correlation analysis was performed on the time-frequency spectrum and the instantaneous frequency curve to obtain the correlation coefficient matrix of the high-voltage variable frequency speed control integrated machine; Based on the correlation coefficient matrix, an insulation defect feature matrix of the high-voltage variable frequency speed control integrated machine is constructed; The winding insulation degradation level assessment module is used to assess the degradation degree of the amplitude variation and distribution pattern of the characteristic frequency components in the insulation defect characteristic matrix, and obtain the winding insulation degradation level of the high-voltage variable frequency speed control integrated machine.

Images