CN114325361A - Transformer tap switch state online monitoring method based on vibration principle - Google Patents
Transformer tap switch state online monitoring method based on vibration principle Download PDFInfo
- Publication number
- CN114325361A CN114325361A CN202111344532.3A CN202111344532A CN114325361A CN 114325361 A CN114325361 A CN 114325361A CN 202111344532 A CN202111344532 A CN 202111344532A CN 114325361 A CN114325361 A CN 114325361A
- Authority
- CN
- China
- Prior art keywords
- vibration
- signal
- tap changer
- transformer
- state
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a transformer tap changer state online monitoring method based on a vibration principle, which comprises the following steps: step 1: the method comprises the steps of carrying out acoustic vibration signal acquisition and preprocessing on a transformer on-load tap changer, determining and verifying acoustic vibration signal sensor type selection and sensor preferred installation mode and position, and improving reliability and accuracy of the sensor; step 2: obtaining a signal envelope line and setting a corresponding threshold value to realize monitoring and feature extraction of a motor driving current signal; and step 3: fault diagnosis and operation state evaluation of the tap changer. The invention can select the tap changer which has heavier load, more frequent operation and more faults as a demonstration application test point, after the test point is successful, the vibration state monitoring of the tap changer is realized, effective guidance and means are provided for operation management personnel, the operation maintenance work intensity and time are reduced, the operation maintenance level and the operation reliability of the transformer substation are improved, and the change of the maintenance mode and the perfection of the state maintenance are promoted.
Description
Technical Field
The invention relates to the technical field of electric power, in particular to a transformer tap changer state online monitoring method based on a vibration principle.
Background
The transformer of the common transformer substation is an indispensable part in a power grid as a key link of power transformation, and the reliable operation of the transformer is significant for ensuring the safety of the power grid. Statistics of the last decade show that the average loss of electricity in each transformer accident reaches millions of kilowatts, the economic loss caused by the transformer accident is thousands of times or even tens of thousands of times of the price of the equipment, and the transformer accident is more caused by the key component of the on-load tap-changer. Therefore, the power operation department has put an urgent need and higher requirements on ensuring the operational reliability of the on-load tap-changer of the transformer.
The on-load tap-changer is the only movable part in the on-load tap-changer and is also one of the key parts. The on-load tap-changer is used for accurately and timely acting, so that the large-amplitude fluctuation of voltage can be reduced and avoided, the safe and reliable operation of a power system is ensured, and the flexibility of power grid dispatching is improved. As the only mechanically operable component of a transformer, a single operation of an on-load tap changer involves a series of actuation events in which mechanical vibration signals are generated in response to contact impact, friction, etc. Typically, these vibration signals can be tested using in vitro sensing. When some fault hidden dangers exist in the on-load tap-changer, vibration signals on the surface of the on-load tap-changer caused by the action of the contact are different from those in a normal state, so that the vibration waveforms of the action processes are measured and recorded, and are analyzed, and compared with normal signals, the operation working condition of the on-load tap-changer can be effectively reflected. The operation condition of the on-load tap-changer is directly related to the operation safety of the on-load tap-changer, and the faults of the on-load tap-changer are increased along with the increase of the application of the on-load tap-changer in a power grid. Therefore, it is necessary to research the on-line monitoring technology of the tap changer.
Disclosure of Invention
The invention aims to provide a transformer tap changer state online monitoring method based on a vibration principle, which improves the operation maintenance level and the operation reliability of a transformer substation, promotes the transformation of a maintenance mode and the perfection of state maintenance, and solves the problems in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme:
the transformer tap changer state on-line monitoring method based on the vibration principle comprises the following steps:
step 1: the method comprises the steps of carrying out acoustic vibration signal acquisition and preprocessing on a transformer on-load tap changer, determining and verifying acoustic vibration signal sensor type selection and sensor preferred installation mode and position, and improving reliability and accuracy of the sensor;
step 2: obtaining a signal envelope line and setting a corresponding threshold value to realize monitoring and feature extraction of a motor driving current signal;
and step 3: fault diagnosis and operation state evaluation of the tap changer.
Furthermore, data of the comprehensive fault diagnosis system is from a sensing test system, multi-sensor monitoring signals are subjected to signal preprocessing and corresponding signal characteristic extraction, logic reasoning is carried out by an analysis system, a vibration signal comprehensive analysis processing program can be called at any time according to needs, reasoning results are subjected to fusion processing, and finally a diagnosis conclusion and information such as reasoning explanation and the like are output together to complete comprehensive fault diagnosis of mechanical performance of the tap changer.
Further, the on-load tap changer acoustic vibration signal acquisition of step 1 comprises the following steps:
the characteristic signal of the tap switch is converted into a voltage signal through a vibration sensor and a current transformer and is output, and after passing through a voltage limiting protection circuit, a low-pass filter circuit and a voltage following signal conditioning circuit, the CPLD controls the time sequence of the A/D chip to complete the processes of channel selection and analog-digital conversion data acquisition and conversion; and after the A/D chip is converted, an interrupt signal is generated, the CPLD sends a reading instruction to the A/D after receiving the interrupt signal, and simultaneously sends an INT signal to inform the DSP of reading the conversion result of the A/D.
Furthermore, when the system is started to collect signals, firstly, default configuration parameters are used for setting memory resources and bus occupied resources of the system, and entering the entry setting of a main program, resetting all registers, setting a broken vector and the like.
Further, the preprocessing of step 1 includes pre-filtering, zero averaging, and eliminating the trend term.
Further, step 2, a driving motor current acquisition hardware system is installed in the electric operating mechanism box, a motor current signal in the action process of the tap switch is detected in real time through a remote software system, and the switching mechanism jam fault diagnosis is carried out on the basis of the motor current signal.
Further, step 3 includes collecting and processing vibration signals in a laboratory, collecting and processing vibration signals on site, and evaluating and correcting the state of the difference degree of the vibration characteristics.
Further, aiming at the collection and processing of laboratory vibration signals, fault signals are collected and processed through a laboratory transformer on-load tap changer, main fault characteristics corresponding to fault types are extracted, and the difference between the fault characteristics and the fault characteristics in a normal state is analyzed.
Further, aiming at field vibration signal acquisition and processing, a large amount of OLTC vibration signals are acquired and analyzed by carrying out signal detection on a large amount of field transformer on-load tap changers, the acquired vibration signals are processed by using a filter to obtain characteristic vibration signals of the vibration signals, vibration characteristics are extracted, and the characteristic vibration signals are compared with the actual characteristic signals of OLTC typical abnormal operation and fault cases for analysis.
Further, for the state evaluation of the vibration feature difference degree, the feature signal is analyzed and described, the characteristics of the OLTC under normal and typical faults are compared according to the obtained information of the peak time, the amplitude and the like of the vibration signal, and an evaluation method based on a support vector machine is used.
Compared with the prior art, the invention has the beneficial effects that:
the invention can select the tap changer which has heavier load, more frequent operation and more faults as a demonstration application test point, and after the test point is successful, the tap changer is popularized and applied to other transformer substations, thereby realizing the monitoring of the vibration state of the tap changer, providing effective guidance and means for operation managers, reducing the working intensity and time of operation and maintenance, improving the operation and maintenance level and the operation reliability of the transformer substations, and promoting the transformation of the maintenance mode and the perfection of state maintenance. The rationalization, standardization and scientization of equipment maintenance are realized, and the method is suitable for the new situation that national economy development requires high quality and high reliability of electric power.
Drawings
FIG. 1 is a block diagram of the present invention;
fig. 2 is a flow chart of the work of the on-load tap-changer on-line monitoring system of the transformer of the invention;
FIG. 3 is a diagram of the hardware design of the signal acquisition unit of the present invention;
FIG. 4 is a signal acquisition flow chart of the present invention;
FIG. 5 is a flow chart of vibration signal acquisition and preprocessing of the present invention;
FIG. 6 is a flow chart of the preferred steps for analyzing the mounting location of the vibration sensor of the present invention;
FIG. 7 is a derivative of the envelope of the signal according to the present invention;
FIG. 8 is a schematic diagram of a support vector machine of the present invention;
FIG. 9 is a schematic diagram of a laboratory difference analysis of the present invention;
FIG. 10 is a diagram illustrating the state evaluation of the degree of difference in vibration characteristics according to the present invention;
fig. 11 is a schematic diagram of the fault diagnosis of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The transformer tap changer state on-line monitoring method based on the vibration principle comprises the following steps:
step 1: the method is characterized in that acoustic vibration signal acquisition and preprocessing are carried out on the transformer on-load tap changer, the acoustic vibration signal sensor type selection and the preferable installation mode and position of the sensor are determined and verified, and the reliability and accuracy of the sensor are improved, and refer to fig. 1.
And the data of the comprehensive fault diagnosis system is derived from the sensing test system. After signal preprocessing and corresponding signal feature extraction are carried out on the multi-sensor monitoring signals, an analysis system carries out logic reasoning, a vibration signal comprehensive analysis processing program can be called at any time according to needs, the reasoning results are fused, and finally a diagnosis conclusion and information such as reasoning explanation and the like are output together to complete comprehensive fault diagnosis of mechanical performance of the tap switch, and refer to fig. 2.
Vibration signal preprocessing principle: the raw vibration signal obtained by data acquisition is often mixed with various noises, non-linearity in the measuring instrument, and the like, and these adverse components may have a great influence on the final analysis result. Therefore, the test data needs to be preprocessed before the analysis process to improve the reliability and authenticity of the vibration signal analysis. The preprocessing includes pre-filtering, zero averaging, and eliminating trend terms.
(1) Pre-filtering
When the signal needs to be smoothed or the unwanted frequency components are suppressed, a filtering method may be employed. To avoid frequency aliasing that occurs because the sampling theorem is not satisfied, a low-pass filter can be used to limit the bandwidth of the original signal while reducing high-frequency noise. The digital filter also counteracts drift and avoids power leakage.
For a digital filter, the signals involved are all discrete signals, and the output sequence y (n) is a discrete convolution of the input sequence x (n) and the impulse response sequence h (n), that is:
in actual calculation, infinite multiple terms and finite terms M cannot be removed, or a feedback form is adopted, and the past output of the filter is reused for calculating the weight of the next stage. Thus, two calculation methods are obtained, and when T is 1, the calculation formula is as follows:
formula (II)
Described are finite impulse response Filters (FIR), and
an infinite impulse response filter (IIR) is described. Conventionally, the weight coefficients h (i) and g (j) are represented by a (i) and b (j), and when N is far larger than M and K in the formula, the formula is paired
And
the system functions of the two filters can be obtained by performing Z transformation as follows:
the impulse response sequence h (k) of the digital filter has infinite length, if a data window w (m) with length N is used to intercept the first N items which can sufficiently reflect the characteristics of the filter, the infinite sequence is changed into a finite sequence, and the digital calculation can be realized on the premise of not using a feedback form, and the interception process is as follows:
hw(m)＝w(m)h(k),k＝-∞,...,0,...∞
n is an odd number, and m is- (N-1)/2
N/2, N being an odd integer, N ═ N/2
In the formula (I), the compound is shown in the specification,
the input-output relationship of the filter at this time is as follows:
because the finite impulse response filter does not adopt a feedback form, namely the current output is irrelevant to the past output and is generally in a non-transmission type structure, the negative sign of the finite impulse response filter is omitted in consideration of the symmetry of the filter, only the factor sequence is considered, and T is equal to 1, the formula
Can be simplified as follows:
it can be seen that the FIR filter characteristics are actually determined by the impulse response sequence h (m), and varying the impulse response sequence can result in different types of filters, such as high-pass, low-pass, band-pass, and band-stop.
(2) Zero equalization
In order to analyze the statistical characteristics of the signal and eliminate the direct current component in the data, zero-mean processing needs to be performed on the signal. Assuming the original signal sequence is x (n), the zero-mean data sequence is:
in the formula
Is the average of the original sequence x (n). The average value may be averaged over all the 1 st to nth samples. If the data is sampled from different segments, averaging and zero mean transformation can be performed during each segment.
(3) Eliminating trend terms
The trend term refers to a linear term or a slowly varying nonlinear component having a period greater than the recording length, which is present in a random signal. The existence of the trend term can cause large errors in the correlation analysis in the time domain and the power spectrum analysis in the frequency domain, and even completely loses the authenticity of the low frequency spectrum. The least square method is usually adopted to eliminate the linear trend term included in the signal, and the specific method is as follows:
assuming the use of a k-th order polynomial vnTo fit the digital signal unThen, there are:
selecting a suitable coefficient b according to the least square principleiSo that v isnAnd unThe sum of squared errors between the two is minimal. The sum of the squared errors is:
in the formula unIs the original signal, vnFitting the signal for a polynomial, T being the sampling period, biIs a polynomial coefficient.
The term in eliminating the linear trend is b0And b1Two terms, i.e. k is 1, wherein b0Is mean value, b1The slope of the linear trend term, then:
this gives a binomial expression: v. ofn＝b0+b1nT。
Structural vibration testing often hopes to accurately acquire linear independent vibration response data as much as possible by using a limited number of sensors, and the larger the response of a measuring point is, the more beneficial the acquisition of signals and the identification of parameters are, so that the installation position of the sensors should be reasonably selected, whether the installation position reasonably and directly influences the quality of the tested signals, and whether the vibration characteristics of the tested object can be accurately reflected. The theory of optimizing the sensor position mainly includes: (1) an effective independent method; (2) an effective independent drive point residual method; (3) mode weighting effectively independent. The main principle of each method is as follows:
1) effective independent law
The main idea of the effective independence method is to select the nodes that contribute most to the independence of the structural target mode shape to maximize the spatial resolution of the target mode shape. The optimization of the sensor position is realized by selecting the point which enables the determinant value of the Fisher information matrix to be maximum, the mathematical expression is shown as follows, phi is the mode shape normalized according to the mass, F is the Fisher information matrix, and E is the effective independent distribution matrix.
F＝φTφ
E＝φF-1φT
Distribution matrix [ E]Is a symmetric power matrix, each diagonal element of which represents the corresponding measurement point-to-mode matrix [ phi ]]The degree of contribution of the rank of (c). EiiThe values are from 0 to 1, 0 representing that it does not contribute to the rank of the target mode shape, and 1 representing that it contributes most to the rank of the target mode shape. In order to make the selected measuring point reflect the target mode shape of the structure better, the measuring point close to 1 should be selected preferentially.
2) Effective independent-drive point residual error method
The effective independence method is a very effective tool for identifying linearly independent measurement modes as much as possible by selecting an optimal measurement point, but does not consider the response of the measurement point, so that some measurement points with low average response may be selected. Considering the difference in the types of the measured responses of the sensors, the average responses of the respective points are classified into the following categories.
Shift average response:
acceleration average response:
weighting the effective independent distribution matrix with the average response of each point to obtain:
ED＝E(i)×ADDOFD(i)
EV＝E(i)×ADDOFV(i)
EA＝E(i)×ADDOFA(i)
the method comprises the steps of selecting a formula to calculate a parameter value Ex according to a sensor type selected by a structural vibration modal test, wherein points with larger Ex indicate that the parameter value Ex has high contribution to the spatial resolution of a target vibration mode and has larger average response, circularly deleting points with the minimum contribution to the determinant value of Ex each time by using a stepwise subtraction method, circularly iterating in such a way until the number of required measurement points remains, wherein the remaining measurement points are the optimal measurement points calculated by an effective independent driving point residual method, the optimal measurement points have the largest contribution to the spatial resolution of the structural target vibration mode, the average response is also larger, the acquisition of signals is facilitated, and the optimal measurement points are selected preferentially.
3) Vibration mode weighted effective independence method
The main idea of the effective independent method is to select a node which has the largest contribution to the independence of the structural target mode shapes so as to maximize the spatial resolution of the target mode shapes, however, the effective independent method considers that the contribution of each target mode shape is the same when selecting the measurement points, however, the correlation degree between some mode shapes of the structure is higher, and the mode shapes are difficult to identify in practice, or some mode shapes may be more interested in the practical mode test, and it is hopeful that the arrangement of the measurement points of some mode shapes of some order or some mode shapes can be improved. Therefore, an effective independent vibration mode weighting method is provided, the effective independent vibration mode weighting method can weight certain order or certain order vibration modes according to modal test requirements, and the weight of the interested vibration mode is improved, so that the spatial resolution of the interested vibration mode is improved, and the interested vibration mode is better identified.
The mode shape weighting effectively independent method weights the mode shape first. Phi is the mode of each order normalized by mass, phirIs the r-th order mode, p is the weighting coefficient for the r-th order mode, and the formula is as follows:
calculating a Fisher information matrix by using the unweighted mode shapes, wherein the formula is as follows:
F＝φTφ
and calculating an effective independent distribution matrix by using the weighted mode shape, wherein the formula is as follows:
EPIs an effective independent distribution matrix of mode shape weighting, similar to the effective independent method, each diagonal element represents the contribution degree of its corresponding measurement point to the mode shape matrix. The higher the diagonal line element is, the greater the contribution of the corresponding measuring point to the target vibration mode is, which is beneficial to the spatial identification of the target vibration mode.
After the optimal installation position of the sensor is obtained, the installation mode of the sensor needs to be considered, the installation mode of the acceleration sensor on the vibration structure is very important for obtaining reliable results, in the vibration measurement process, the installation of the acceleration sensor needs to enable the preset measurement direction to be coincident with the main sensitivity axis, the sensor needs to be in good contact with a measured object, if sliding is generated in the horizontal direction or contact is separated in the vertical direction, the test result is seriously distorted, and the reliability of the test data is reduced, so that the test data cannot be used. In addition, the natural frequency of the spring-mass system formed by the mounting connection stiffness of the acceleration sensor and the mass of the sensor becomes the mounting resonance frequency, which is often lower than the upper limit of the use frequency of the sensor, so that the mounting resonance frequencies of the sensors in different mounting modes are different. Therefore, what method is used to mount the acceleration sensor on the measuring point has a significant influence on the measurement accuracy and range.
Collecting acoustic vibration signals of the on-load tap-changer: the characteristic signal of the tap switch is converted into a voltage signal through a vibration sensor and a current transformer and is output, and after passing through a voltage limiting protection circuit, a low-pass filter circuit and a voltage following signal conditioning circuit, the CPLD controls the time sequence of the A/D chip to complete the processes of data acquisition and conversion, such as channel selection, analog-to-digital conversion and the like; after the conversion of the a/D chip is finished, an interrupt signal is generated, the CPLD receives the interrupt signal and then sends a read instruction to the a/D, and simultaneously sends an INT signal to notify the DSP to read the conversion result of the a/D, see fig. 3.
When the system is started to collect signals, firstly, default configuration parameters are used for setting memory resources and bus occupied resources of the system, and the default configuration parameters enter the entry setting of a main program, all registers are cleared, the setting of a broken vector is carried out, and other preparation work before the main program runs is carried out. And entering the main program operation of system software, and controlling the time sequence of the A/D chip to finish the process of signal sampling. And judging whether the sampling frequency of the A/D chip reaches the set frequency through a counter, if not, continuing sampling, and otherwise, performing A/D data conversion. When the A/D chip finishes the data conversion work, the effective interrupt signal with low level is output, and whether the A/D chip finishes the data conversion is determined by judging whether the signal of the interrupt output port is low level. When the interrupt signal port outputs a low level signal (namely, the A/D finishes data conversion), the A/D reads data, the digital quantity after 14-bit conversion is stored in the SRAM, and finally the data in the SARM is read. The signal acquisition control flow chart is shown in fig. 4.
The vibration signal acquisition and preprocessing flow is as shown in fig. 5, and the acquired tap changer original vibration signal is processed by a preprocessing means, an unnecessary frequency component is suppressed by a pre-filtering method, a direct current component in data is eliminated by a zero-averaging method, and a linear term or a nonlinear component with a period greater than a recording length is eliminated by an elimination trend term method. And (3) extracting characteristic parameters of the preprocessed signals by using algorithms such as short-time Fourier transform and the like, simulating typical faults through a laboratory, and performing difference analysis on characteristic values of the typical faults and characteristic values of normal conditions.
Based on laboratory test data, the position of the sensor is optimized by adopting an effective independent algorithm, the optimization method of the effective independent algorithm can be carried out by adopting a direct method, a step-by-step accumulation method and a step-by-step elimination method, the position of the sensor is optimized by selecting a proper algorithm by comparing the influence of various search optimization algorithms on the selection of the position of the sensor, the position of a measuring point is evaluated by modal correlation, measuring point energy or a Fisher information matrix, and finally the optimal position for mounting the sensor is obtained. The influence of the installation mode of the sensor on the data acquisition signal is researched, the two methods of adhesion and permanent magnet adsorption are respectively tested and contrastively analyzed, so that a reasonable installation mode of the sensor is selected, the type selection, the installation position and the mode of the sensor are finally tested and verified through a field test, the vibration sensor is tested in a X, Y, Z three-dimensional space multipoint mode, and the test analysis steps are shown in fig. 6.
Step 2: and (3) solving a signal envelope line and setting a corresponding threshold value to realize monitoring and feature extraction of the motor driving current signal.
Analysis theory of current signals of the driving motor: if the amplitude change of the current of the driving motor exceeds a certain range, the abnormal condition is shown in the switching process of the switch. For monitoring and characteristic extraction of motor driving current signals, a method of obtaining signal envelope lines and setting corresponding threshold values is adopted.
The envelope curve of the signal is not complex in theory, if the envelope curve is drawn manually, the envelope curve of the corresponding signal can be obtained only by selecting and rejecting the brain and connecting the corresponding extreme points, but in a digital signal processing system, the envelope curve is not simple. There are generally three methods of extracting the envelope of a signal: hilbert amplitude demodulation, demodulation-filtering, and high-pass absolute value demodulation. The real part of the analytic signal obtained by Hilbert transformation is a signal, the imaginary part of the analytic signal is Hilbert transformation of the signal, and the amplitude of the analytic signal is an envelope of the signal: the detection-filtering method is that the original signal is subjected to detection processing, then zero equalization is carried out, band-pass filtering with the factory as the center frequency is set, and envelope signals with the factory as the main component can be obtained; the high-pass absolute value method is to perform high-pass filtering on a zero-mean time domain signal, and then perform low-pass filtering after taking an absolute value, wherein the selection of low-pass frequency determines the frequency component of an envelope signal. The high-pass absolute value method is used for enveloping the central line of the signal, the detection-filtering method is used for enveloping the central line of the positive half cycle of the signal, and the demodulation amplitude spectrum obtained by the two methods is not the real enveloping amplitude. The envelope demodulated by the Hilbert method is an envelope of absolute values of signals, the demodulated amplitude of the envelope represents a real signal envelope, and in a digital signal processing system, Hilbert transform can be conveniently realized by means of fast Fourier transform, so that the envelope of current signals is extracted by the Hilbert transform method.
The drive motor current acquisition hardware system can be installed in the electric operating mechanism box, a motor current signal in the action process of the tap switch is detected in real time through a remote software system, and the switching mechanism jam fault diagnosis is carried out on the basis of the motor current signal. The specific derivation process of extracting the envelope of the current signal by using the Hilbert transform method and solving the envelope of the signal is shown in fig. 7. After an envelope curve of a motor driving current signal is correctly obtained, a characteristic value average current, a difference value between the minimum current and the maximum current in a steady state and an operation length are extracted. The average current (average current in a steady state) is mainly used for judging the problem of the load increase of the motor. The difference between the minimum current and the maximum current in a steady state mainly judges the lubrication problems of the main gear, the bearing and the transmission mechanism. The length of operation (total time between initial current and end of current) is primarily determinative of the up/down shift characteristic, odd/even shift characteristic.
And step 3: fault diagnosis and operation state evaluation of the tap changer.
(1) Fault diagnosis algorithm
1) Short time Fourier transform
Tap changer vibration signals are non-stationary signals and need to be represented using a two-dimensional union of the time domain and the frequency domain to be accurately described.
The short-time Fourier transform, also known as windowed Fourier transform, was proposed by Gabor in 1946. The basic idea is as follows: the signal is divided into a number of small time intervals, and each time interval is analyzed by Fourier transform to determine the frequency at which the interval exists, for the purpose of time-frequency localization. The expression of the short-time Fourier transform is:
in the formula (I), the compound is shown in the specification,
is the conjugate of g (t), g (t) is a time limit function, also called a window function, and plays a time limit role; e.g. of the type-iωtAnd plays a role of frequency limitation. g (t) and e-iωtThe combination may be time-frequency localized. Fgf (ω, τ) roughly reflects the relative content of f (t) its signal component at time τ at frequency ω.
As defined above, windowed Fourier transform transforms achieve a certain degree of time-frequency localization, and are suitable for deterministic stationary signals. Once the window function is determined and the size and shape of the window are fixed, then the time and frequency resolution is unity. Let ω be m ω0，τ＝nτ0，ω0、τ0Respectively sampling step lengths of a frequency domain and a time domain; when the window function g (t) is not changed, then m ω is calculated for different frequency components0The sampling steps in the time domain are each τ0。
If the window function satisfies the following equation:
the signal can be completely reconstructed with its STFT transform, and its generalized inverse transform can be written as:
2) discrete short time Fourier transform
In practical application, F needs to be treatedSTFTx(T, F) discretizing is performed, for which samples are taken at equally spaced time-frequency grid points (mT, nF) on the time-frequency plane, where T and F are the sampling intervals of the time and frequency variables, respectively. Let x (k) denote a discrete form of signal x (t), k being 0,1,2, …, N-1; where m, N is 0,1,2, …, N-1, and N is the total number of sample points, the discrete form of the STFT transform is:
it can be seen that the STFT transform results in 1 two-dimensional complex matrix, and fast computation can be achieved using FFT transform, whose magnitude matrix can be expressed as:
A(m,n)＝|FSTFT(m,n)|
the obtained signal STFT is used to transform an amplitude matrix, the rows of which correspond to sampling time points, the columns of which correspond to frequency values, and the elements of the matrix are corresponding spectrum amplitudes, it should be noted that the actual frequency corresponding to the frequency coordinate value N at this time is equal to N/(NT), N is the total number of sampling points, and the value range N of the variable N is 0,1,2, …, N/2-1. Thus, in the amplitude matrix a (m, n), each row vector is a frequency spectrum at a corresponding time, and each column vector is a time distribution of each frequency spectrum.
(2) State evaluation method
The state evaluation method needs to process different fault signals of the tap changer to obtain an evaluation criterion, and finally, the state of the tap changer corresponding to the signal can be obtained by judging the characteristics of a certain field signal and the difference degree between the characteristics and the threshold value in the evaluation criterion. The evaluation method is based on signal processing, the evaluation methods are multiple, and the evaluation method based on the support vector machine is selected.
The principle of applying the support vector machine to the state evaluation of the tap changer is as follows: the method takes each state quantity of the tap changer as input, and then adjusts the relevant parameters of the support vector machine to carry out learning and testing, so that different samples can be input to obtain corresponding output values. After a certain number of samples are trained by adopting the method, an evaluation system can be obtained, and then the weight coefficients of all indexes are stored in a training model, so that the nonlinear regression can be realized, as shown in fig. 8.
The vibration signal of the tap changer directly reflects the characteristic parameters of certain faults and is the basis of state evaluation. The deep research on the failure mechanism and the clear understanding of the internal relation between the state quantities corresponding to the vibration signals are necessary preconditions for accurate state evaluation. The state evaluation process is a subjective process, and the accuracy of the evaluation result is greatly influenced by human factors. In consideration of the influence of each vibration characteristic signal on the health condition of the tap changer, the weight coefficients of each state quantity need to be assigned, and are corrected and perfected through a large amount of data. Meanwhile, in the evaluation process, it is difficult to avoid erroneous judgment or missed judgment of the fault signal, and the judgment criterion of the state evaluation needs to be corrected and perfected, please refer to fig. 9.
1. Laboratory vibration signal acquisition and processing
And fault signals are collected and processed through the on-load tap changer of the transformer in the laboratory, main fault characteristics corresponding to the fault types are extracted, and the difference between the main fault characteristics and the normal state is analyzed.
2. In-situ vibration signal acquisition and processing
A large amount of OLTC vibration signals are collected and analyzed by carrying out signal detection on a large amount of on-load tap-changers of the field transformers. And processing the acquired vibration signal by using a filter to obtain a characteristic vibration signal of the vibration signal. And extracting vibration characteristics, and comparing and analyzing the vibration characteristics with characteristic signals of actual OLTC typical abnormal operation and fault cases. Different typical index quantities can reflect normal, abnormal and fault operation states of different OLTC switching actions, and the actual operation state and long-term operation trend of the OLTC can be quantitatively reflected by adopting a proper data processing method. Data normalization processing can be adopted to preliminarily judge the relevant operating state of the equipment according to the magnitude of the relative value.
3. State evaluation of degree of difference in vibration characteristics
Analyzing and describing the characteristic signals, comparing the characteristics of the OLTC under normal and typical faults according to the obtained information of the peak time, the amplitude and the like of the vibration signals, obtaining an evaluation criterion of the OLTC under the typical working state by using the difference degree of the signal characteristics corresponding to each fault and determining a difference degree threshold value by using an evaluation method based on a support vector machine and the like and referring to FIG. 10. The operating state of the tap changer within the respective threshold values is determined using the threshold values of the degree of difference in the evaluation criterion.
4. Perfect correction
The on-load tap-changer mechanical characteristic on-line monitoring system based on the vibration analysis method is applied and developed, the detection and the diagnosis of the on-load tap-changer mechanical characteristic under the field working condition are carried out, the reason for generating errors of the detection device is discussed according to the field test result, corresponding improvement measures are provided, and the on-load tap-changer mechanical characteristic on-line monitoring system and a corresponding fault diagnosis algorithm are perfected.
The method comprehensively judges the problems of the on-load tap-changer gear-up/down characteristic or odd/even characteristic, motor load, lubrication problem, synchronization, braking, contact wear and the like by detecting and analyzing the characteristic values of the on-load tap-changer such as the operation length (total time between initial current and current end), the average current (total time between initial current and current end), the difference value of the minimum current and the maximum current in a steady state, the gear-shifting time length (time difference value between gear switching end and current end), the high-frequency/low-frequency maximum amplitude and the like, thereby realizing the expected target of tap-changer fault type judgment, wherein the fault diagnosis scheme is shown as the following figure 11.
The invention is based on the vibration principle, collects the vibration signal of the tap changer and the current signal of the driving motor through the high-precision vibration acceleration sensor and the current transformer, and carries out characteristic recognition on the signals by using a mathematical analysis method, thereby realizing online monitoring of the operation state of the tap changer and realizing accurate judgment of the parallel synchronization consistency of the tap changer. The project is based on a maintenance technology taking reliability and preventability as centers, realizes the identification of latent early signs of faults of the tap changer, judges fault positions, fault severity and development tendency, evaluates running states and makes a reliable maintenance plan, thereby supporting the running maintenance of the tap changer, effectively prolonging the service life of the tap changer and improving the safe running level of a power transmission and transformation system.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be able to cover the technical solutions and the inventive concepts of the present invention within the technical scope of the present invention.
Claims (10)
1. The transformer tap changer state online monitoring method based on the vibration principle is characterized by comprising the following steps of:
step 1: the method comprises the steps of carrying out acoustic vibration signal acquisition and preprocessing on a transformer on-load tap changer, determining and verifying acoustic vibration signal sensor type selection and sensor preferred installation mode and position, and improving reliability and accuracy of the sensor;
step 2: obtaining a signal envelope line and setting a corresponding threshold value to realize monitoring and feature extraction of a motor driving current signal;
and step 3: fault diagnosis and operation state evaluation of the tap changer.
2. The on-line transformer tap changer state monitoring method based on the vibration principle as claimed in claim 1, wherein the data of the comprehensive fault diagnosis system is derived from a sensing test system, the multi-sensor monitoring signals are subjected to signal preprocessing and corresponding signal feature extraction, then logical reasoning is performed by the analysis system, a vibration signal comprehensive analysis processing program can be called at any time according to needs, then the reasoning results are subjected to fusion processing, and finally, the diagnosis conclusion and the reasoning interpretation and other information are output together to complete the comprehensive fault diagnosis of the tap changer mechanical performance.
3. The vibration-based on-line monitoring method for the state of the tap changer of the transformer according to claim 1, wherein the step 1 of collecting the acoustic vibration signal of the on-load tap changer comprises the following steps:
the characteristic signal of the tap switch is converted into a voltage signal through a vibration sensor and a current transformer and is output, and after passing through a voltage limiting protection circuit, a low-pass filter circuit and a voltage following signal conditioning circuit, the CPLD controls the time sequence of the A/D chip to complete the processes of channel selection and analog-digital conversion data acquisition and conversion; and after the A/D chip is converted, an interrupt signal is generated, the CPLD sends a reading instruction to the A/D after receiving the interrupt signal, and simultaneously sends an INT signal to inform the DSP of reading the conversion result of the A/D.
4. The vibration-principle-based on-line monitoring method for the state of the tap changer of the transformer according to claim 3, wherein when the system is started to collect signals, default configuration parameters are firstly used to set the memory resources and the bus occupied resources of the system, and the main program is entered into the entry setting, all registers are cleared, the setting of the break vector is performed, and other preparation works before the main program is run.
5. The method for vibration-based on-line monitoring of the condition of the tap changer of the transformer according to claim 1, wherein the preprocessing of step 1 comprises pre-filtering, zero averaging and trend elimination terms.
6. The vibration-principle-based on-line monitoring method for the state of the tap changer of the transformer according to claim 1, wherein in the step 2, a driving motor current acquisition hardware system is installed in an electric operating mechanism box, a motor current signal in the action process of the tap changer is detected in real time through a remote software system, and the fault diagnosis of the jamming of the switching mechanism is carried out based on the motor current signal.
7. The vibration-principle-based online monitoring method for the state of the tap changer of the transformer according to claim 1, wherein the step 3 comprises the steps of laboratory vibration signal acquisition and processing, field vibration signal acquisition and processing, and state evaluation and correction perfection of the degree of difference of vibration characteristics.
8. The vibration-principle-based on-line monitoring method for the state of the tap changer of the transformer according to claim 7, wherein for the collection and processing of laboratory vibration signals, the collection and processing of fault signals are performed through the on-load tap changer of the transformer in the laboratory, main fault characteristics corresponding to fault types are extracted, and the difference between the main fault characteristics and the normal state is analyzed.
9. The method as claimed in claim 7, wherein for on-line monitoring of the on-load tap changer status of the transformer based on the vibration principle, the on-load tap changer of a large number of on-load tap changers of the transformer on the field is acquired and processed, a large number of OLTC vibration signals are acquired and analyzed, the acquired vibration signals are processed by using a filter to obtain the characteristic vibration signals thereof, the vibration characteristics are extracted, and the characteristic vibration signals are compared and analyzed with the characteristic signals of the actual OLTC typical abnormal operation and fault cases.
10. The on-line monitoring method for transformer tap changer state based on vibration principle as claimed in claim 7, characterized in that, for the state evaluation of vibration feature difference, the feature signal is analyzed and described, and the characteristics of OLTC under normal and typical fault are compared according to the obtained information of vibration signal peak time, amplitude, etc. by using the evaluation method based on support vector machine.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111344532.3A CN114325361A (en) | 2021-11-15 | 2021-11-15 | Transformer tap switch state online monitoring method based on vibration principle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111344532.3A CN114325361A (en) | 2021-11-15 | 2021-11-15 | Transformer tap switch state online monitoring method based on vibration principle |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114325361A true CN114325361A (en) | 2022-04-12 |
Family
ID=81045558
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111344532.3A Pending CN114325361A (en) | 2021-11-15 | 2021-11-15 | Transformer tap switch state online monitoring method based on vibration principle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114325361A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116577716A (en) * | 2023-07-06 | 2023-08-11 | 西安高压电器研究院股份有限公司 | Current sensor vibration characteristic testing method, related equipment and related system |
| CN116956203A (en) * | 2023-09-21 | 2023-10-27 | 山东和兑智能科技有限公司 | Method and system for measuring action characteristics of tapping switch of transformer |
| CN117150421A (en) * | 2023-11-01 | 2023-12-01 | 江苏沙洲电气有限公司 | Novel low-voltage switch cabinet data monitoring method and system |
-
2021
- 2021-11-15 CN CN202111344532.3A patent/CN114325361A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116577716A (en) * | 2023-07-06 | 2023-08-11 | 西安高压电器研究院股份有限公司 | Current sensor vibration characteristic testing method, related equipment and related system |
| CN116577716B (en) * | 2023-07-06 | 2023-10-20 | 西安高压电器研究院股份有限公司 | Current sensor vibration characteristic testing method, related equipment and related system |
| CN116956203A (en) * | 2023-09-21 | 2023-10-27 | 山东和兑智能科技有限公司 | Method and system for measuring action characteristics of tapping switch of transformer |
| CN116956203B (en) * | 2023-09-21 | 2023-12-15 | 山东和兑智能科技有限公司 | Method and system for measuring action characteristics of tapping switch of transformer |
| CN117150421A (en) * | 2023-11-01 | 2023-12-01 | 江苏沙洲电气有限公司 | Novel low-voltage switch cabinet data monitoring method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114325361A (en) | Transformer tap switch state online monitoring method based on vibration principle | |
| Kim et al. | Fault diagnosis of a power transformer using an improved frequency-response analysis | |
| CN110728257B (en) | Transformer winding fault monitoring method based on vibration gray level image | |
| CN109581270B (en) | Method for rapidly identifying CVT fault based on wavelet multi-scale transformation | |
| WO2013185405A1 (en) | Power transformer iron core looseness fault diagnosis and preliminary fault location system and method | |
| CN108802535B (en) | Screening method, main interference source identification method and device, server and storage medium | |
| CN109443717B (en) | On-load tap-changer mechanical fault on-line monitoring method | |
| CN104251953A (en) | Diagnosis method and device for loosening faults of on-load tap switch contact of power transformer | |
| CN114152825A (en) | Fault diagnosis method and device of transformer and fault diagnosis system of transformer | |
| CN110991481A (en) | High-voltage shunt reactor internal loosening fault diagnosis method based on cross wavelet transformation | |
| CN113866684A (en) | Distribution transformer fault diagnosis method based on hybrid sampling and cost sensitivity | |
| Ma et al. | A novel time-frequency slice extraction method for target recognition and local enhancement of non-stationary signal features | |
| Yu et al. | A new method to select frequency band for vibration signal demodulation and condition estimation of rolling bearings | |
| Chen et al. | Noise-robust adaptive feature mode decomposition method for accurate feature extraction in rotating machinery fault diagnosis | |
| CN110929673A (en) | Transformer winding vibration signal identification method based on ITD (inverse discrete cosine transformation) permutation entropy and CGWO-SVM (Carrier-support vector machine) | |
| Tse et al. | Enhanced eigenvector algorithm for recovering multiple sources of vibration signals in machine fault diagnosis | |
| Zhou et al. | Multi-node feature learning network based on maximum spectral harmonics-to-noise ratio deconvolution for machine condition monitoring | |
| Wan et al. | Adaptive asymmetric real Laplace wavelet filtering and its application on rolling bearing early fault diagnosis | |
| CN113032915A (en) | Method and device for evaluating health state of main shaft bearing of manufacturing equipment | |
| Cui et al. | Egram based SVD method for gear fault diagnosis | |
| CN115343579B (en) | Power grid fault analysis method and device and electronic equipment | |
| CN110645887A (en) | Winding deformation judgment method based on short-circuit reactance | |
| CN108593095A (en) | A kind of Vibration Fault Signal acquiring and processing method of converter power transformer tap switch | |
| CN115932699A (en) | Multi-dimensional diagnosis method and system for signal sampling channel | |
| Ma et al. | Envelope demodulation method based on SET for fault diagnosis of rolling bearings under variable speed |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |