CN-121978591-A - Transformer mechanical state in-situ sensing method based on spatial magnetic field modal evolution

CN121978591ACN 121978591 ACN121978591 ACN 121978591ACN-121978591-A

Abstract

The invention provides a primary taste sensing method for a mechanical state of a transformer based on space magnetic field modal evolution, which comprises the steps of extracting a steady-state magnetic field component and a vibration magnetic field component in the transformer, acquiring a first detection feature and a second detection feature according to the steady-state magnetic field component, diagnosing whether the magnetic circuit state of a transformer core is normal or not based on the first detection feature and the second detection feature, and diagnosing whether the axial vibration mode and the radial vibration mode of a transformer winding are normal or not according to the vibration magnetic field component. According to the invention, through fusing the spatial modal characteristics of the internal magnetic field of the transformer with the vibration frequency characteristics, the external measurement signals of the transformer are strongly correlated with the changes of the internal mechanical state, so that the fault types of the winding and the iron core can be effectively distinguished, the fault area can be preliminarily positioned, and the diagnosis accuracy is greatly improved.

Inventors

ZHAO NANNAN
WU YUHUI
LIU XINCHUN
LI SIWEN
LIAO BIN

Assignees

吉安伊戈尔磁电科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260109

Claims (10)

1. A transformer mechanical state primary taste perception method based on spatial magnetic field modal evolution is characterized in that: Step A, arranging a plurality of sensors in a transformer according to an array arrangement principle, wherein each sensor collects time-varying magnetic field signals of the transformer and gathers the time-varying magnetic field signals into an original space-time data set; step B, extracting steady-state magnetic field components and vibration magnetic field components based on the original space-time data set; Extracting steady-state magnetic field components comprises extracting the total harmonic distortion rate of the magnetic field, and performing data processing on the total harmonic distortion rate of the magnetic field to construct a total harmonic distortion rate data sequence of the magnetic field; Extracting the vibration magnetic field component comprises constructing a space snapshot for reflecting the space distribution of the vibration magnetic field at any moment, and constructing a space-time matrix based on the space snapshots at a plurality of moments; Step C, acquiring a first detection feature and a second detection feature according to a magnetic field total harmonic distortion rate data sequence, and diagnosing whether the magnetic circuit state of the transformer core is normal or not based on the first detection feature and the second detection feature; when the magnetic circuit of the transformer core is abnormal, triggering the abnormal early warning of the core, and not executing the step D; D, when the magnetic circuit state of the transformer core is normal, executing the step D; step D, extracting a spatial vibration mode from the time matrix based on the intrinsic orthogonal decomposition POD, calculating the mode vibration frequency and constructing a vibration profile based on the spatial vibration mode, and diagnosing whether the axial and radial vibration modes of the transformer winding are normal or not according to the mode vibration frequency and the vibration profile; and E, outputting the diagnosis results of the magnetic circuit state of the transformer core and the axial and radial vibration modes of the transformer winding.
2. The primary taste perception method of the mechanical state of the transformer based on the modal evolution of the spatial magnetic field according to claim 1, which is characterized in that: in the step A, a plurality of sensors are distributed in the transformer according to an array distribution principle, and the method comprises the following steps: Three vertical measuring lines are arranged on the side surface of the wall of the transformer oil tank along the height direction, the three vertical measuring lines respectively correspond to the axial projection positions of the high-voltage winding, the low-voltage winding and the iron core, and a plurality of high-precision triaxial magneto-resistance sensors are distributed on each vertical measuring line at unequal intervals; The number of the high-precision three-axis magnetic resistance sensors arranged at the end part and the middle part of the corresponding winding is more than that of the high-precision three-axis magnetic resistance sensors arranged at other positions of the winding.
3. The primary taste perception method of the mechanical state of the transformer based on the modal evolution of the spatial magnetic field according to claim 1, which is characterized in that: in the step a, each sensor collects time-varying magnetic field signals of the transformer and sums the time-varying magnetic field signals into an original spatiotemporal dataset, including: Each sensor performs continuous acquisition operation at a sampling frequency higher than the highest vibration power frequency of the transformer winding and the iron core, so as to obtain a time-varying magnetic field signal of each sensor , Representing a time-varying magnetic field signal acquired by an ith sensor at time t, the time-varying magnetic field signal comprising three orthogonal components of a magnetic field vector , M represents the total number of sensors, T represents the time length of the sampling time window; and integrating time-varying magnetic field signals acquired by the M sensors at all moments in a sampling time window into an original space-time data set.
4. A method for primary perception of mechanical state of a transformer based on modal evolution of a spatial magnetic field according to claim 3, wherein the method comprises the steps of: extracting the total harmonic distortion of the magnetic field, comprising: for all time-varying magnetic field signals within each period of each sensor in the original spatio-temporal data set Performing Fourier transformation to obtain fundamental wave amplitude and each subharmonic amplitude, and calculating the total harmonic distortion rate of the magnetic field according to a formula I: -equation one; Wherein, the Indicating the total harmonic distortion of the magnetic field of the ith sensor during one cycle, Representing the fundamental amplitude of the ith sensor in one cycle, Representing the H-th harmonic amplitude of the ith sensor in one period, H representing the highest harmonic order in one period, ; The total harmonic distortion of the magnetic field of each sensor is obtained in a plurality of continuous periods.
5. The primary taste perception method for the mechanical state of the transformer based on the modal evolution of the spatial magnetic field according to claim 4, wherein the primary taste perception method is characterized in that: in the step B, performing data processing on the total harmonic distortion rate of the magnetic field to construct a total harmonic distortion rate data sequence of the magnetic field, including: setting the sizes of the observation window and the basic windows to obtain N basic windows, , Indicating the size of the observation window and, Representing the size of the base window; each basic window comprises a total harmonic distortion rate of the magnetic field of each sensor in a plurality of continuous periods; Removing the total harmonic distortion rate of the magnetic field belonging to the instantaneous interference field point in a basic window for the total harmonic distortion rate of the magnetic field of all sensors in the basic window, and taking the average value of the total harmonic distortion rates of other residual magnetic fields in the basic window as the representative total harmonic distortion rate of the magnetic field of the basic window; acquiring representative total harmonic distortion of magnetic field of all basic windows to form a data sequence of total harmonic distortion of magnetic field of an observation window , Representative magnetic field total harmonic distortion rate representing a first fundamental window, Representative field total harmonic distortion rate representing a second fundamental window, Representative field total harmonic distortion rate for the nth fundamental window is represented.
6. The primary taste perception method for the mechanical state of the transformer based on the modal evolution of the spatial magnetic field according to claim 5, wherein the primary taste perception method is characterized in that: The step C comprises the following steps: calculating a first detection feature according to a formula II, and calculating a second detection feature according to a formula III; -formula two; -formula three; Representing a first detection characteristic, namely a mean shift of the total harmonic distortion rate of the magnetic field, Magnetic field total harmonic distortion rate data sequence representing current observation window Is used for the average value of (a), Representing a long-term reference mean value of a total harmonic distortion rate data sequence of a healthy magnetic field, which is obtained by measuring a transformer in a healthy state in advance; representing a second detection feature, namely a dispersion ratio of the total harmonic distortion rate of the magnetic field, Magnetic field total harmonic distortion rate data sequence representing current observation window Is set in the standard deviation of (2), Representing a long-term reference standard deviation of a total harmonic distortion rate data sequence of a healthy magnetic field, which is obtained by measuring a transformer in a healthy state in advance; Setting an average value abnormality threshold And a dispersion anomaly threshold ; Determining a first diagnostic condition: Determining a second diagnostic condition Determining a third diagnostic condition , And Representing a mean value anomaly lower limit value and a mean value anomaly upper limit value; When the second diagnosis condition and/or the third diagnosis condition are met, the magnetic circuit of the transformer core is abnormal, the abnormal early warning of the transformer core is triggered, and the step D is not executed; when the second diagnosis condition is not met and the third diagnosis condition is not met, the magnetic circuit of the transformer core is normal, and the step D is executed; And when the first diagnosis condition is met, judging that the detection is invalid.
7. A method for primary perception of mechanical state of a transformer based on modal evolution of a spatial magnetic field according to claim 3, wherein the method comprises the steps of: in the step B, extracting the vibration magnetic field component includes constructing a spatial snapshot for reflecting the spatial distribution of the vibration magnetic field at any one time, constructing a space-time matrix based on the spatial snapshots at a plurality of times, including: Time-varying magnetic field signals for each sensor in the raw spatiotemporal dataset Digital bandpass filtering to preserve the core vibration signal of each sensor , Representing a time-varying magnetic field signal which is noise filtered out by digital bandpass filtering Then rearranging the time t according to the time continuity to obtain a time point; for any moment The vibration signal amplitude of the same orthogonal component measured by all M sensors is formed into M-dimensional vector , , Indicating that after digital band-pass filtering, it is in Normal component of Mth sensor at time T represents matrix transposition, M-dimensional vector Recording as a space snapshot; continuously taking n time points , ,..., Is used to determine the spatial snapshot of (1), Arranged in time sequence as A space-time matrix X of dimensions, 。
8. The primary taste perception method of the mechanical state of the transformer based on the modal evolution of the spatial magnetic field of claim 7, which is characterized in that: In the step D, extracting a spatial vibration mode from the time matrix based on the eigen-orthogonal decomposition POD includes: calculating a covariance matrix C on the space matrix X according to the fourth formula; -formula four; performing eigenvalue decomposition on the covariance matrix C according to a formula five; -formula five; represents the j-th POD mode, i.e., j spatial vibration modes, expressed as A feature vector of the dimension, wherein each element in the feature vector corresponds to the vibration amplitude weight of one sensor; Represents the j-th feature value representing the j-th POD mode The degree of contribution in the total vibration energy, The larger the POD mode corresponding to the POD mode The more important it is that, ; All characteristic values are calculated The first k characteristic values are taken according to the sequence from big to small, so as to obtain k dominant POD modes 。
9. The primary taste perception method of the mechanical state of the transformer based on the modal evolution of the spatial magnetic field of claim 8, which is characterized in that: In said step D, calculating modal vibration frequencies based on the spatial vibration modes and constructing a vibration profile comprises: for a dominant POD modality, calculating a modality time coefficient of the dominant POD modality according to formula six; -formula six; representing dominant POD modalities Is represented by the modal time coefficient of (1) In the vibration signal of moment, the j-th dominant POD mode has a law that the contribution degree changes along with time; representing dominant POD modalities Is a transpose of (2); Performing a Fourier transform calculation on the modal time coefficient of a dominant POD mode to obtain the modal vibration frequency of the dominant POD mode ; For dominant POD modality Interpolation and visualization is performed to a two-dimensional projection of the transformer tank to form a vibration profile.
10. The primary taste perception method of the mechanical state of the transformer based on the modal evolution of the spatial magnetic field according to claim 9, wherein the primary taste perception method is characterized in that: diagnosing whether the axial and radial vibration modes of the transformer winding are normal according to the mode vibration frequency and the vibration sectional view, comprising: setting an axial first diagnosis condition and an axial second diagnosis condition, wherein the axial first diagnosis condition comprises that for one dominant POD mode Its modal vibration frequency Lower frequency limit less than the reference low frequency of the vibration of the health mode The second diagnostic condition includes, for a dominant POD modality The phase difference between the amplitude of the high-voltage side of the corresponding winding and the amplitude of the low-voltage side of the winding is larger than a first preset phase difference, and the amplitude of the high-voltage side of the winding and the amplitude of the low-voltage side of the winding are shown to be in opposite phase according to a vibration profile; When a dominant POD modality When the first diagnosis condition and the second diagnosis condition are met, the abnormal axial vibration mode of the transformer winding is judged, the axial compression force of the transformer winding is loose, and the radial vibration mode of the transformer winding is not required to be judged; When a dominant POD modality When the axial first diagnosis condition or the axial second diagnosis condition is met, judging results are not made on whether the axial vibration mode of the transformer winding is abnormal or not, and the radial vibration mode of the transformer winding is continuously judged; When a dominant POD modality If the first diagnosis condition or the second diagnosis condition is not met, judging that the axial vibration mode of the transformer winding is normal, and continuously judging the radial vibration mode of the transformer winding; Setting a radial core diagnostic condition, a radial assist first diagnostic condition, and a radial assist second diagnostic condition, the radial core diagnostic condition comprising, for a dominant POD modality Its modal vibration frequency Reference intermediate frequency range of healthy mode vibration frequency The phase difference between the amplitudes of the same level of the corresponding windings is smaller than or equal to a second preset phase difference, and the amplitudes of the same level of the windings are shown to be in phase according to the vibration profile, the radial auxiliary first diagnosis condition comprises that for a dominant POD mode The second diagnosis condition of radial assistance comprises the following steps of for one dominant POD mode Its characteristic value Greater than the health characteristic value ; When a dominant POD modality When the radial core diagnosis condition is met, the radial vibration mode abnormality of the transformer winding is directly judged, and the radial deformation of the transformer winding is expressed; When a dominant POD modality When the radial core diagnosis condition is not met but the radial auxiliary first diagnosis condition and the radial auxiliary second diagnosis condition are met, judging that the radial vibration mode of the transformer winding is abnormal, wherein the radial vibration mode is expressed as radial deformation of the transformer winding; When a dominant POD modality When the radial core diagnosis condition, the radial auxiliary first diagnosis condition and the radial auxiliary second diagnosis condition are not met, judging that the radial vibration mode of the transformer winding is normal, and continuously judging whether the transformer winding has a local high-frequency mode or not; Setting a high frequency mode first diagnosis condition and a high frequency mode second diagnosis condition, wherein the high frequency mode first diagnosis condition comprises that for one dominant POD mode Its modal vibration frequency Is greater than the upper limit of the reference high frequency of the vibration frequency of the health mode The second diagnosis condition of the high-frequency mode comprises the following steps of for a dominant POD mode The energy concentration coefficient is larger than the healthy energy concentration coefficient; When a dominant POD modality When the first diagnosis condition of the high-frequency mode and the second diagnosis condition of the high-frequency mode are met, judging that the local high-frequency mode exists in the transformer winding, and the local high-frequency mode appears as local mechanical fault of the winding; When a dominant POD modality And when the first diagnosis condition of the high-frequency mode and/or the second diagnosis condition of the high-frequency mode are not met, judging that the local high-frequency mode does not exist in the transformer winding.

Description

Transformer mechanical state in-situ sensing method based on spatial magnetic field modal evolution Technical Field The invention belongs to the technical field of transformer winding and iron core fault detection, and particularly relates to a transformer mechanical state in-situ sensing method based on spatial magnetic field modal evolution. Background The transformer is used as core equipment of the power grid, and the mechanical integrity of the internal winding and the iron core is the fundamental point for ensuring safe operation. The existing state monitoring technology mainly has the following defects: 1. Indirection and hysteresis a widely used oil chromatography analysis judges faults by detecting insulating oil breakdown gases, which is a typical "post hoc" diagnosis. Only after the insulating material is cracked due to overheat or discharge, the insulating material can be detected, and mechanical faults such as loosening and deformation of the winding can not be early warned; 2. The method is insensitive to mechanical states, and although an electrical test (such as short-circuit impedance and frequency response analysis) can reflect winding deformation, the method is usually carried out in a power failure mode and cannot realize continuous online monitoring. Although the vibration analysis method is online, the vibration analysis method is easy to be interfered by background noise, and the sensor is usually arranged on the base of the oil tank and is insensitive to slight mechanical state change inside the winding; 3. the prior method mostly provides integral state evaluation, is difficult to accurately position whether faults occur on a high-voltage side, a low-voltage side or an iron core, and lacks the spatial resolution capability on the internal state of a transformer; None of these prior art techniques essentially enable direct, early and in situ sensing of the mechanical state of the windings and cores of the transformer. Disclosure of Invention Aiming at the problems of the background technology, the invention provides a transformer mechanical state primary taste sensing method based on spatial magnetic field modal evolution. To achieve the purpose, the invention adopts the following technical scheme: a transformer mechanical state primary taste sensing method based on spatial magnetic field modal evolution comprises the following steps: Step A, arranging a plurality of sensors in a transformer according to an array arrangement principle, wherein each sensor collects time-varying magnetic field signals of the transformer and gathers the time-varying magnetic field signals into an original space-time data set; step B, extracting steady-state magnetic field components and vibration magnetic field components based on the original space-time data set; Extracting steady-state magnetic field components comprises extracting the total harmonic distortion rate of the magnetic field, and performing data processing on the total harmonic distortion rate of the magnetic field to construct a total harmonic distortion rate data sequence of the magnetic field; Extracting the vibration magnetic field component comprises constructing a space snapshot for reflecting the space distribution of the vibration magnetic field at any moment, and constructing a space-time matrix based on the space snapshots at a plurality of moments; Step C, acquiring a first detection feature and a second detection feature according to a magnetic field total harmonic distortion rate data sequence, and diagnosing whether the magnetic circuit state of the transformer core is normal or not based on the first detection feature and the second detection feature; when the magnetic circuit of the transformer core is abnormal, triggering the abnormal early warning of the core, and not executing the step D; D, when the magnetic circuit state of the transformer core is normal, executing the step D; step D, extracting a spatial vibration mode from the time matrix based on the intrinsic orthogonal decomposition POD, calculating the mode vibration frequency and constructing a vibration profile based on the spatial vibration mode, and diagnosing whether the axial and radial vibration modes of the transformer winding are normal or not according to the mode vibration frequency and the vibration profile; and E, outputting the diagnosis results of the magnetic circuit state of the transformer core and the axial and radial vibration modes of the transformer winding. Preferably, in the step a, a plurality of sensors are arranged inside the transformer according to an array arrangement principle, including: Three vertical measuring lines are arranged on the side surface of the wall of the transformer oil tank along the height direction, the three vertical measuring lines respectively correspond to the axial projection positions of the high-voltage winding, the low-voltage winding and the iron core, and a plurality of high-precision triaxial magneto-r