Extra-high voltage direct current converter transformer hidden fault diagnosis system

Abstract

The invention, which relates to the technical field of the transformer, discloses an extra-high-voltage DC converter transformer hidden fault diagnosis system comprising a multi-dimensional acquisition end, a feature decoupling end and a cross-domain cooperation end. According to the invention, the deformation unit builds a model by means of direct-current magnetic bias current and magnetic flux density, the heat dissipation unit analyzes by means of Pearson correlation coefficients, pipeline faults are effectively identified, early warning is realized through mean value calculation and change monitoring, oil temperature stability is guaranteed, the insulation unit extracts features through wavelet packet multi-scale decomposition, faults are diagnosed according to energy abrupt change, and insulation hidden dangers are accurately captured. A mechanical unit constructs a delay coordinate matrix through winding vibration signals, the track separation rate is quantified by means of the maximum Lyapunov exponent, hidden faults of a mechanical structure are recognized, all the units cooperate with multiple dimensions to guarantee equipment stability, and fault diagnosis and operation and maintenance efficiency is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of transformers, in particular to a kind of implicit fault diagnosis system of UHV DC converter transformer. BACKGROUND

[0002] As the core equipment of DC power transmission system, UHV DC converter transformer undertakes the key functions such as AC-DC power conversion and voltage level conversion, and its safe and stable operation is crucial to guarantee power grid transmission capacity and maintain power system reliability. With the wide application of UHV power transmission technology, converter transformer faces complex working conditions (such as DC bias, harmonic interference, environmental magnetic field coupling, etc.), and implicit faults (such as core magnetic property degradation, winding slight deformation, early insulation degradation, etc.) gradually become hidden dangers threatening equipment safety.

[0003] At present, the traditional monitoring means in the prior art mainly relies on single parameter threshold alarm, which is difficult to effectively identify early implicit faults. Such faults have progressive and coupled characteristics, for example, vibration anomalies caused by core magnetic bias coexist with insulation degradation, and are affected by strong electromagnetic interference and multi-physical field coupling, so conventional diagnosis methods are prone to miss report or misdiagnosis.

[0004] Therefore, the present application proposes a kind of implicit fault diagnosis system of UHV DC converter transformer to solve the above problems. SUMMARY

[0005] The main purpose of the present application is to provide an implicit fault diagnosis system of UHV DC converter transformer to solve the problems raised in the above background.

[0006] To achieve the above purpose, the technical scheme adopted by the present application is as follows: an implicit fault diagnosis system of UHV DC converter transformer, the system comprises a multi-dimensional acquisition end, a feature decoupling end and a cross-domain collaboration end;

[0007] The multi-dimensional acquisition end is used to collect multi-modal data of transformer body and external environment in real time through Internet of Things sensors and perform preprocessing;

[0008] The feature decoupling end calculates the core power loss main frequency offset based on DC bias current and magnetic flux density parameters, calculates the Pearson correlation coefficient of DC bias current and oil temperature, performs wavelet packet decomposition on converter valve voltage waveform and extracts frequency band energy, detects insulation defects according to energy mutation, finally calculates the maximum Lyapunov exponent of winding vibration signal, and performs comprehensive diagnosis and early warning of multiple types of implicit faults;

[0009] The cross-domain cooperative end builds a characteristic correlation matrix of vibration and temperature, when the maximum singular value of the matrix exceeds a set threshold, performs secondary diagnosis confirmation, and based on the partial discharge ultrahigh frequency signal and the environmental noise signal, calculates the characteristic frequency energy ratio of the discharge and the noise, combines the signal singular value decomposition to separate the fault component, when the energy ratio exceeds the threshold and the principal component energy proportion exceeds the threshold, the secondary positioning insulation deterioration fault is performed.

[0010] Preferably, the multi-dimensional acquisition end includes a body unit, an external unit and a preprocessing unit.

[0011] The body unit is used to acquire the direct current bias current data of the transformer through the current sensor, the magnetic flux density parameter of the transformer through the magnetic flux density sensor, the temperature data of the transformer oil through the oil temperature sensor, the voltage waveform data of the converter valve through the voltage sensor, the vibration signal data of the winding through the vibration sensor, the temperature data of the hot spot position of the transformer through the temperature sensor and the ultrahigh frequency signal data of the internal partial discharge of the transformer through the partial discharge ultrahigh frequency sensor.

[0012] The external unit is used to acquire the noise signal data of the transformer operating environment through the noise sensor.

[0013] The preprocessing unit is used to perform denoising, abnormal value replacement and normalization processing on the collected multi-modal data.

[0014] Preferably, the characteristic decoupling end includes a deformation unit, a heat dissipation unit, an insulation unit and a mechanical unit.

[0015] Preferably, the deformation unit builds an association model based on the direct current bias current data and the magnetic flux density parameter, and then determines the proportional coefficient in the model by means of the finite element algorithm, uses the model to decouple and analyze the deviation of the core caused by the bias and the associated vibration characteristics, and identifies the hidden fault of the core according to the decoupling result.

[0016] Preferably, the heat dissipation unit uses the Pearson correlation coefficient analysis method to calculate the linear correlation degree between the preprocessed direct current bias current data and the oil temperature sampling data, and identifies the pipeline fault of the transformer by analyzing the value and change of the correlation coefficient.

[0017] Preferably, the insulation unit adopts wavelet packet decomposition means to perform multi-scale decomposition processing on the voltage waveform signal according to the voltage waveform of the converter valve, obtains signal characteristics of different levels and different frequency bands, calculates the energy value of the corresponding frequency band at each decomposition level, calculates the energy difference of adjacent time points by monitoring the change of the frequency band energy, and outputs the corresponding fault diagnosis information according to the energy difference.

[0018] Preferably, the mechanical unit is based on the collected winding vibration signal, constructs a delay coordinate matrix of the vibration signal according to the delay time and the embedding dimension, converts the one-dimensional vibration signal into a high-dimensional space sequence, analyzes the constructed delay coordinate matrix, calculates the maximum Lyapunov exponent, and quantifies the stability of the winding mechanical structure by the index.

[0019] Preferably, the cross-domain cooperative terminal comprises a distinguishing unit, a first cooperative unit, a second cooperative unit and a warning unit, the distinguishing unit is used for receiving fault judgment results of the variable unit, the heat dissipation unit, the insulation unit and the mechanical unit, and outputting corresponding operations.

[0020] Preferably, the first cooperative unit is based on the amplitude and frequency of the winding vibration signal and the temperature rise rate data of the hot spot area of the transformer, constructs a feature correlation matrix, and includes the covariance of the vibration amplitude and the temperature rise rate, the covariance of the vibration frequency and the temperature rise rate, the covariance of the vibration amplitude and the frequency itself, and the variance information of the temperature rise rate, finally monitors the singular value of the matrix, compares the singular value with the set threshold, and outputs the judgment result.

[0021] Preferably, the second cooperative unit is based on the partial discharge ultrahigh frequency signal and the environmental noise signal, extracts the energy of the corresponding characteristic frequency through time-frequency transformation, calculates the energy ratio of the two, then performs singular value decomposition on the signal, separates out the discharge fault related component, and calculates the energy ratio and the energy proportion of the fault component after decomposition, when the energy ratio and the main energy of the fault component both exceed the set threshold, it is determined that there is an insulation deterioration fault.

[0022] The warning unit is based on the results of the first cooperative unit and the second cooperative unit to output corresponding warning information.

[0023] The present application has the following advantages:

[0024] 1. In the present application, the variable unit constructs a model based on the direct current bias magnetic current and the magnetic flux density, fits the coefficients through the finite element algorithm, decouples the core bias deviation and the vibration characteristics, accurately identifies the hidden faults, the heat dissipation unit is based on the direct current bias magnetic current and the oil temperature data, uses the Pearson correlation coefficient analysis to effectively identify the pipeline fault, realizes early warning through mean value calculation and change monitoring, ensures the stability of the oil temperature, the insulation unit is aimed at the voltage waveform of the converter valve, extracts the features through wavelet packet multi-scale decomposition, diagnoses the fault according to the energy mutation, accurately captures the insulation hidden danger, the mechanical unit constructs a delay coordinate matrix from the winding vibration signal, quantifies the trajectory separation rate by the maximum Lyapunov index, identifies the mechanical structure hidden fault, and each unit cooperates in multiple dimensions to ensure the stability of the equipment and improve the fault diagnosis and operation efficiency.

[0025] 2. In the present application, the distinguishing unit intelligently schedules, the first cooperative unit fuses vibration and temperature multi-physical quantities, and excavates coupled abnormalities by using characteristic correlation matrix and singular value decomposition, solving the problem of missed judgment under multi-physical field interference; the second cooperative unit integrates partial discharge ultra-high frequency and noise signals, and through time-frequency transformation, energy ratio analysis and fault component separation, it accurately captures the implicit characteristics of insulation degradation, the early warning unit responds in stages, double faults trigger emergency shutdown, and single-dimensional abnormalities start special re-measurement, through multi-source data cooperation and cross-physical quantity coupling analysis, it effectively identifies progressive and coupled faults such as core magnetic bias and insulation degradation, solves the problem of missed report and misjudgment under strong interference, and improves the coverage and accuracy of implicit fault diagnosis. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 It is a framework diagram of the implicit fault diagnosis system of the extra-high voltage direct current converter transformer of the present application.

[0027] Figure 2 It is a flowchart of the implicit fault diagnosis system of the extra-high voltage direct current converter transformer of the present application. DETAILED DESCRIPTION

[0028] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0029] Please refer to Figure 1 and Figure 2 , the present application provides a technical solution: an implicit fault diagnosis system for an extra-high voltage direct current converter transformer, which comprises a multi-dimensional acquisition end, a feature decoupling end and a cross-domain cooperation end.

[0030] The multi-dimensional acquisition end is used for real-time acquisition of multi-modal data of the transformer body and the outside through Internet of Things sensors, and pre-processing.

[0031] The feature decoupling end calculates the core loss main frequency offset of the iron core based on the direct current magnetic bias current and the magnetic flux density parameter, calculates the Pearson correlation coefficient of the direct current magnetic bias current and the oil temperature, performs wavelet packet decomposition on the voltage waveform of the converter valve and extracts the frequency band energy, detects insulation defects according to energy mutation, and finally calculates the maximum Lyapunov index of the winding vibration signal to comprehensively diagnose and warn multiple types of implicit faults.

[0032] The cross-domain cooperative end builds a characteristic correlation matrix of vibration and temperature, when the maximum singular value of the matrix exceeds a set threshold, secondary diagnosis confirmation is performed, and based on the partial discharge ultrahigh frequency signal and the environmental noise signal, the characteristic frequency energy ratio of the discharge and the noise is calculated, the signal singular value decomposition is separated, and when the energy ratio exceeds the threshold and the principal component energy ratio exceeds the threshold, the insulation deterioration fault is located.

[0033] Specifically, the system first acquires the direct current bias magnetic current, magnetic flux density, oil temperature, vibration signal and other intrinsic parameters and environmental noise and other external parameters through the multi-dimensional acquisition end, forms standardized data after denoising and normalization processing, the characteristic decoupling end performs main frequency shift analysis on the core parameters to detect the magnetic property deterioration, identifies the abnormal heat dissipation through the correlation coefficient of the oil temperature and the bias magnetic current, detects the insulation material frequency band energy mutation through wavelet packet decomposition, and evaluates the winding mechanical stability through the maximum Lyapunov exponent, the cross-domain cooperative end builds a correlation matrix of the vibration signal and the temperature data, verifies the fault correlation through singular value analysis, and performs secondary confirmation of the insulation state in combination with the energy ratio of the partial discharge signal and the noise.

[0034] The multi-dimensional acquisition end includes an intrinsic unit, an external unit and a preprocessing unit;

[0035] The intrinsic unit is used for acquiring the direct current bias magnetic current data of the transformer through a current sensor, acquiring the magnetic flux density parameters of the transformer through a magnetic flux density sensor, acquiring the temperature data of the transformer oil through an oil temperature sensor, acquiring the voltage waveform data of the converter valve through a voltage sensor, acquiring the vibration signal data of the winding through a vibration sensor, acquiring the temperature data of the hot spot position of the transformer through a temperature sensor and acquiring the ultrahigh frequency signal data of the internal partial discharge of the transformer through a partial discharge ultrahigh frequency sensor;

[0036] Specifically, in the data acquisition process, the sampling time is strictly aligned.

[0037] The external unit is used for acquiring the noise signal data of the operating environment of the transformer through a noise sensor;

[0038] The preprocessing unit is used for denoising, abnormal value replacement and normalization processing on the acquired multi-modal data.

[0039] The characteristic decoupling end includes a deformation unit, a heat dissipation unit, an insulation unit and a mechanical unit.

[0040] The deformation unit builds an association model of the direct current bias magnetic current data and the magnetic flux density parameters, and then determines the proportional coefficient in the model by means of a finite element algorithm, decouples the deviation of the core caused by bias magnetization and the associated vibration characteristics by using the model, and identifies the hidden fault of the core according to the decoupling result.

[0041] Specifically, the association model is:

[0042] wherein, Δf represents the core power frequency main frequency offset, k represents the proportional coefficient, B dc represents the DC bias magnetic flux density, B ac represents the AC magnetic flux density, I dc represents the DC bias magnetic current;

[0043] According to the finite element algorithm, the transformer core is modeled, and the core vibration and main frequency offset under different DC bias magnetic currents, DC bias magnetic flux densities, and AC magnetic flux densities are calculated, and the value of k is fitted;

[0044] The real-time collected DC bias magnetic current, DC bias magnetic flux density, and AC magnetic flux density are substituted into the correlation model, and the core power frequency main frequency offset is calculated;

[0045] Based on the actual vibration monitoring core power frequency main frequency offset and the calculated core power frequency main frequency offset, the decoupling core bias deviation is calculated, and the formula is as follows:

[0046]

[0047] wherein, Δf1 represents the actual vibration monitoring core power frequency main frequency offset, Δf represents the calculated core power frequency main frequency offset, and ΔB represents the decoupling core bias deviation;

[0048] When |ΔB| is greater than the bias threshold, it is determined that the core has a hidden fault, wherein the bias threshold is calibrated based on historical fault data, otherwise it is determined to be in a normal state.

[0049] The heat dissipation unit calculates the linear correlation degree between the pre-processed DC bias magnetic current data and the oil temperature sampling data based on the Pearson correlation coefficient analysis method, and identifies the transformer pipeline fault by analyzing the value and change of the correlation coefficient.

[0050] Specifically, after the mean value calculation of the DC bias magnetic current data and the oil temperature sampling data, the linear correlation degree between them is calculated, and the formula is as follows:

[0051]

[0052] wherein, represents the Pearson correlation coefficient between the current and the oil temperature, x i represents the DC bias magnetic current at the i-th moment, y i represents the oil temperature at the i-th moment, represents the mean value of the DC bias magnetic current data, represents the mean value of the oil temperature sampling data;

[0053] If |r| is less than the correlation threshold or the decrease in r within one hour (the decrease in one hour = the correlation coefficient of the previous window one hour ago - the correlation coefficient of the current window) is greater than or equal to the decrease threshold, then it is judged as a hidden pipeline fault; where the correlation threshold and the decrease threshold are both calibrated based on historical fault data.

[0054] The insulation unit uses wavelet packet decomposition to decompose the voltage waveform of the converter valve into multiple scales, obtaining signal characteristics at different levels and frequency bands. It calculates the energy value of the corresponding frequency band at each decomposition level, calculates the energy difference between adjacent moments by monitoring the changes in frequency band energy, analyzes the energy difference, and outputs corresponding fault diagnosis information.

[0055] Specifically, the voltage waveform signal undergoes multi-scale decomposition processing, and the calculation formula is as follows:

[0056]

[0057] Where u(t) represents the voltage waveform signal, and J represents the number of decomposition layers. K W represents the number of frequency bands in the j-th layer. j,k (t) represents the signal characteristics of the k-th frequency band in the j-th layer;

[0058] For the signal characteristics of the k-th frequency band in the j-th layer, calculate the energy value of its frequency band using the following formula:

[0059]

[0060] Among them, E j,k J represents the energy value of the k-th frequency band in the j-th layer, J represents the number of decomposition layers, k represents the number of frequency bands in the j-th layer, N represents the total number of sampling points of the discrete signal, and i represents the sampling point number.

[0061] The energy difference between adjacent time points is calculated using the following formula:

[0062] ΔE j,k (t)=E j,k (t)-E j,k (t-Δt);

[0063] Where, ΔE j,k (t) represents the energy difference between adjacent moments, and Δt represents the time interval. When |ΔE j,k If (t) is greater than twice the average energy difference under normal operating conditions, it indicates an energy mutation and a fault in the corresponding frequency band.

[0064] Based on the collected winding vibration signal, the mechanical unit constructs a delay coordinate matrix of the vibration signal according to the delay time and embedding dimension, converting the one-dimensional vibration signal into a high-dimensional spatial sequence. Then, the constructed delay coordinate matrix is ​​analyzed to calculate the maximum Lyapunov exponent, which is used to quantify the stability of the winding mechanical structure.

[0065] Specifically, the delay coordinate matrix is:

[0066]

[0067] Where x(t) represents the time-domain sequence of the vibration signal, t represents the time variable, τ represents the delay time, m represents the embedding dimension, and N represents the total number of sampling points of the discrete signal;

[0068] Analyze adjacent trajectories in the delay coordinate matrix and calculate the maximum Lyapunov exponent using the following formula:

[0069]

[0070] Where, λ max The maximum Lyapunov exponent quantifies the rate of separation between adjacent trajectories, and t represents the evolution time. M X represents the number of trajectory pairs. i (t) and X j (t) represents the adjacent trajectory at time t, X i (0) and X j (0) represents the adjacent trajectory at the initial time;

[0071] The separation threshold is determined by calibrating historical data, when λ max If the value is greater than the separation threshold, it indicates a hidden fault in the mechanical structure; otherwise, it indicates normal operation.

[0072] The cross-domain collaborative terminal includes a differentiation unit, a first collaborative unit, a second collaborative unit, and an early warning unit. The differentiation unit is used to receive the fault results from the deformation unit, the heat dissipation unit, the insulation unit, and the mechanical unit, and output the corresponding operation.

[0073] If any one of the deformation unit, heat dissipation unit, and mechanical unit is determined to be faulty, the system enters the first coordination unit; if the insulation unit is determined to be faulty, the system enters the second coordination unit.

[0074] The first collaborative unit constructs a feature correlation matrix based on the amplitude and frequency of the winding vibration signal and the temperature rise rate data of the transformer hot spot area. It incorporates the covariance of vibration amplitude and temperature rise rate, the covariance of vibration frequency and temperature rise rate, the covariance of vibration amplitude and its own frequency, and the variance of temperature rise rate. Finally, it monitors the singular values ​​of the matrix, compares the singular values ​​with the set threshold, and outputs the judgment result.

[0075] Specifically, the constructed feature correlation matrix is:

[0076]

[0077] wherein A represents a winding vibration amplitude sequence, f represents a winding vibration frequency sequence, represents an oil temperature rise rate sequence, Cov(X, Y) represents a covariance of X and Y, represents a covariance of the oil temperature rise rate;

[0078] The singular value decomposition of the feature correlation matrix is performed by Python numpy, all singular values are traversed, the largest singular value is selected and compared with a threshold value of the historical normal working condition, and if the largest singular value is greater than the threshold value of the historical normal working condition, it is judged that there is a cooperative fault.

[0079] The second cooperative unit extracts the energy of the corresponding characteristic frequency based on the partial discharge ultrahigh frequency signal and the environmental noise signal through time-frequency transformation, calculates the energy ratio of the two, and then performs singular value decomposition on the signal to separate the discharge fault related components, and calculates the energy ratio and the energy proportion of the decomposition fault components, when the energy ratio and the main energy of the fault components both exceed the set threshold value, it is determined that there is an insulation deterioration fault;

[0080] Specifically, the preprocessed partial discharge ultrahigh frequency signal and the environmental noise signal are subjected to short-time Fourier transform to obtain the time-frequency domain matrix of the partial discharge signal and the environmental noise signal, then the energy of the local environmental noise signal and the local discharge signal at the corresponding characteristic frequency is extracted, and the energy ratio of the two is calculated, and the calculation formula is as follows:

[0081]

[0082] wherein R 比值 represents the energy ratio, E 噪声 represents the energy of the extracted local environmental noise signal at the corresponding characteristic frequency, E 放电 represents the energy of the local discharge signal at the corresponding characteristic frequency;

[0083] At the same time, the time is taken as the row, the frequency is taken as the column, and the corresponding energy value is taken as the matrix value to construct the signal matrix, and the singular value decomposition is performed through SVD, the first two largest singular values are retained, the fault component is reconstructed, and the fault component proportion is calculated:

[0084]

[0085] The energy ratio R 比值 and the fault component proportion P are compared with the corresponding energy threshold value and the component threshold value, and if the energy ratio R 比值And the fault component proportion P is greater than the corresponding energy threshold and component threshold, then it is judged as insulation deterioration fault.

[0086] The early warning unit corresponds to the early warning information based on the first cooperative unit and the second cooperative unit result;

[0087] When the variable unit, the heat dissipation unit and the mechanical unit judge as fault, and the first cooperative unit also judges as cooperative fault, or the insulation unit and the second cooperative unit both judge as fault, trigger emergency stop, and remote early warning;

[0088] When the variable unit, the heat dissipation unit and the mechanical unit judge as fault, and the first cooperative unit judges as normal, or the insulation unit judges as fault but the second cooperative unit judges as normal, trigger unit level fault special diagnosis process, automatically dispatch high-precision sensor, and carry out fixed point retest to abnormal unit.

[0089] In the application, a kind of special high voltage direct current converter transformer implicit fault diagnosis system, body unit is collected by multiple sensors, comprehensive acquisition electrical, magnetic, thermal, mechanical and insulation characteristics and other multi-modal data;External unit collects environmental noise, supplements interference information, preprocessing unit carries out denoising, abnormal value replacement and normalization, guarantees data quality, strictly aligns sampling time, ensures time sequence consistency, provides comprehensive, accurate, synchronous data support for subsequent diagnosis, builds the data cornerstone of fault diagnosis;

[0090] Variable unit constructs correlation model with the help of direct current magnetic bias current and magnetic flux density parameters, and then combines finite element algorithm to fit proportion coefficient, decouples the deviation of core caused by magnetic bias and related vibration characteristics, on the one hand, can accurately identify core implicit fault, through the constructed correlation model and decoupling algorithm, subtle abnormalities caused by core magnetic bias can be captured in advance, to avoid fault expansion and ensure stable operation of transformer;On the other hand, based on the comparison of core power frequency main frequency deviation obtained by actual vibration monitoring and calculation, the core magnetic bias deviation is decoupled, so that the fault judgment is more scientific and accurate, and clear fault identification basis is provided for operation and maintenance personnel.

[0091] Based on the preprocessed direct current magnetic bias current and oil temperature sampling data, the heat dissipation unit uses Pearson correlation coefficient analysis to calculate the linear correlation degree of the two, which can effectively identify transformer pipeline faults, discover heat dissipation abnormalities caused by pipeline blockage, leakage and other reasons in time, ensure stable transformer oil temperature, avoid insulation damage and other serious faults caused by overheating, and at the same time, mean value calculation and correlation coefficient change analysis are carried out on the data, which can realize early warning of pipeline implicit fault, calibrate threshold based on historical fault data, make fault identification more in line with actual operation scene, improve fault prediction ability, and ensure that transformer operates in reasonable temperature environment;

[0092] The insulation unit adopts wavelet packet decomposition for multi-scale processing of the converter valve voltage waveform, can accurately extract signal features of different levels and frequency bands through multi-scale decomposition, comprehensively captures subtle changes in the voltage waveform, provides rich basis for insulation defect detection, can effectively identify hidden faults such as insulation aging and partial discharge, calculates the energy of each frequency band and the energy difference between adjacent time points, judges the fault according to the energy mutation, makes the insulation fault diagnosis more targeted and sensitive, can discover the fault at the initial stage, avoid serious consequences such as large-scale power failure caused by insulation fault, and output clear fault diagnosis information through quantitative energy change, help operation and maintenance personnel quickly locate insulation problems;

[0093] The mechanical unit constructs a delay coordinate matrix based on the winding vibration signal, calculates the maximum Lyapunov index, converts the one-dimensional vibration signal into a high-dimensional space sequence, can more comprehensively mine the mechanical structure information contained in the signal, capture the complex characteristics of winding vibration, accurately reflect the stability of the mechanical structure, and the maximum Lyapunov index can quantize the separation rate of adjacent trajectories, effectively identify hidden faults of the mechanical structure such as winding loosening and deformation, and early warning of mechanical fault risk, through historical data calibration of the separation threshold, the fault judgment standard is clear, and more serious problems such as winding damage caused by mechanical failure are avoided.

[0094] The distinguishing unit intelligently schedules, the first cooperative unit fuses vibration and temperature multi-physical quantities, uses feature correlation matrix and singular value decomposition to mine coupled abnormalities and solve the problem of missed judgment under multi-physical field interference; the second cooperative unit integrates partial discharge ultra-high frequency and noise signals, through time-frequency transformation, energy ratio analysis and fault component separation, accurately captures the hidden characteristics of insulation degradation, the early warning unit responds in stages, double faults trigger emergency shutdown, single-dimensional abnormalities start special retesting, through multi-source data cooperation and cross-physical quantity coupling analysis, effectively identify progressive and coupled faults such as core magnetic bias and insulation degradation, solve the problem of missed report and misjudgment under strong interference, and improve the coverage and accuracy of hidden fault diagnosis.

[0095] It should be noted that in this paper, relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between the entities or operations. Moreover, the term "includes" "contains" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0096] While embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely exemplary of the principles and application of the present application. Numerous modifications and adaptions can be effected without departing from the spirit and scope of the present application, which is not limited to the exact construction and arrangement described. It is intended, therefore, to cover all modifications and adaptions that fall within the scope of the claims and their equivalents.

Claims

1. A system for diagnosing an implicit fault of an extra-high voltage direct current converter transformer, characterized in that, The system comprises a multi-dimensional acquisition end, a feature decoupling end and a cross-domain coordination end. The multi-dimensional acquisition end is configured to collect multi-modal data of a transformer body and external environment in real time through Internet of Things sensors and perform preprocessing. The feature decoupling end is configured to calculate a core loss main frequency offset of a DC bias magnetic current and a magnetic flux density parameter, calculate a Pearson correlation coefficient of the DC bias magnetic current and oil temperature, perform wavelet packet decomposition on a converter valve voltage waveform and extract frequency band energy, detect insulation defects according to energy mutation, and finally calculate a maximum Lyapunov index of winding vibration signals to comprehensively diagnose and warn multiple types of hidden faults. The cross-domain coordination end is configured to build a feature correlation matrix of vibration and temperature, perform secondary diagnosis and confirmation when a maximum singular value of the matrix exceeds a set threshold, calculate a feature frequency energy ratio of a partial discharge and noise based on a partial discharge ultrahigh frequency signal and an environmental noise signal, separate fault components through singular value decomposition of the signal, and perform secondary positioning of insulation degradation faults when the energy ratio exceeds a threshold and a principal component energy ratio exceeds a threshold.

2. The system for diagnosing an implicit fault of an EHVDC converter transformer according to claim 1, wherein The multi-dimensional acquisition end comprises a body unit, an external unit and a preprocessing unit. The body unit is configured to collect DC bias magnetic current data of the transformer through a current sensor, magnetic flux density parameters of the transformer through a magnetic flux density sensor, temperature data of transformer oil through an oil temperature sensor, voltage waveform data of the converter valve through a voltage sensor, vibration signal data of the winding through a vibration sensor, temperature data of a hot spot position of the transformer through a temperature sensor, and ultrahigh frequency signal data of internal partial discharge of the transformer through a partial discharge ultrahigh frequency sensor. The external unit is configured to collect noise signal data of the transformer operating environment through a noise sensor. The preprocessing unit is configured to perform denoising, abnormal value replacement and normalization processing on the collected multi-modal data.

3. The system for diagnosing an implicit fault of an EHVDC converter transformer according to claim 1, wherein The feature decoupling end comprises a deformation unit, a heat dissipation unit, an insulation unit and a mechanical unit.

4. The system for diagnosing an implicit fault of an EHVDC converter transformer according to claim 3, characterized in that, The deformation unit is configured to build an association model of the DC bias magnetic current data and the magnetic flux density parameters, determine a proportional coefficient in the model by means of a finite element algorithm, decouple the deviation of the core caused by the bias magnetic and the associated vibration features by using the model, and identify the core hidden faults according to the decoupling results.

5. The system for diagnosing an implicit fault of an EHVDC converter transformer according to claim 3, wherein The heat dissipation unit is configured to calculate the linear correlation degree between the preprocessed DC bias magnetic current data and the oil temperature sampling data by using a Pearson correlation coefficient analysis method, and identify the transformer pipeline faults by analyzing the numerical value and change of the correlation coefficient.

6. The system for diagnosing an implicit fault of an EHVDC converter transformer according to claim 3, wherein The insulation unit is configured to perform multi-scale decomposition processing on the voltage waveform signal by using wavelet packet decomposition means, obtain signal features of different levels and different frequency bands, calculate energy values of corresponding frequency bands at each decomposition level, calculate the energy difference between adjacent time points by monitoring the change of the frequency band energy, analyze the energy difference, and output corresponding fault diagnosis information.

7. The system for diagnosing an implicit fault of an EHVDC converter transformer according to claim 3, wherein The mechanical unit is based on the collected winding vibration signal, according to the delay time and the embedding dimension, constructs the delay coordinate matrix of the vibration signal, converts the one-dimensional vibration signal into a high-dimensional space sequence, and then analyzes the constructed delay coordinate matrix, calculates the maximum Lyapunov index, and quantifies the stability of the winding mechanical structure by the index.

8. The system for diagnosing an implicit fault of an EHVDC converter transformer according to claim 1, wherein The cross-domain cooperative end includes a distinguishing unit, a first cooperative unit, a second cooperative unit and a warning unit. The distinguishing unit is used for receiving fault judgment results of the variable unit, the heat dissipation unit, the insulation unit and the mechanical unit, and outputting corresponding operations.

9. The system for diagnosing an implicit fault of an EHVDC converter transformer according to claim 8, wherein, The first cooperative unit is based on the amplitude and frequency of the winding vibration signal, and the temperature rise rate data of the hot spot area of the transformer, constructs a feature correlation matrix, and includes the covariance of the vibration amplitude and the temperature rise rate, the covariance of the vibration frequency and the temperature rise rate, the covariance of the vibration amplitude and the frequency itself, and the variance information of the temperature rise rate, finally monitors the singular value of the matrix, compares the singular value with the set threshold value, and outputs the judgment result.

10. The system for diagnosing an implicit fault of an EHVDC converter transformer according to claim 8, wherein, The second cooperative unit is based on the partial discharge ultrahigh frequency signal and the environmental noise signal, extracts the energy of the corresponding characteristic frequency through time-frequency transformation, calculates the energy ratio of the two, then performs singular value decomposition on the signal, separates out the discharge fault related components, and calculates the energy ratio and the energy proportion of the decomposed fault components, when the energy ratio and the main energy of the fault component both exceed the set threshold value, it is determined that there is an insulation deterioration fault. The warning unit is based on the results of the first cooperative unit and the second cooperative unit to output corresponding warning information.

Images