CN112201260B - Transformer running state online detection method based on voiceprint recognition - Google Patents
Transformer running state online detection method based on voiceprint recognition Download PDFInfo
- Publication number
- CN112201260B CN112201260B CN202010930602.2A CN202010930602A CN112201260B CN 112201260 B CN112201260 B CN 112201260B CN 202010930602 A CN202010930602 A CN 202010930602A CN 112201260 B CN112201260 B CN 112201260B
- Authority
- CN
- China
- Prior art keywords
- transformer
- comprehensive evaluation
- evaluation index
- voiceprint
- value
- 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.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 54
- 238000011156 evaluation Methods 0.000 claims abstract description 51
- 238000000034 method Methods 0.000 claims abstract description 26
- 238000012545 processing Methods 0.000 claims abstract description 24
- 239000013598 vector Substances 0.000 claims abstract description 22
- 238000009826 distribution Methods 0.000 claims abstract description 15
- 238000009432 framing Methods 0.000 claims abstract description 15
- 230000002159 abnormal effect Effects 0.000 claims abstract description 12
- 238000010606 normalization Methods 0.000 claims description 13
- 239000011159 matrix material Substances 0.000 claims description 10
- 238000001228 spectrum Methods 0.000 claims description 4
- 239000000126 substance Substances 0.000 claims description 4
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 238000012549 training Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 5
- 238000012360 testing method Methods 0.000 description 5
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical group [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000004804 winding Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000001816 cooling Methods 0.000 description 2
- 238000000354 decomposition reaction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 230000005236 sound signal Effects 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004587 chromatography analysis Methods 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 235000021474 generally recognized As safe (food) Nutrition 0.000 description 1
- 235000021473 generally recognized as safe (food ingredients) Nutrition 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000010998 test method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L17/00—Speaker identification or verification techniques
- G10L17/26—Recognition of special voice characteristics, e.g. for use in lie detectors; Recognition of animal voices
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L17/00—Speaker identification or verification techniques
- G10L17/02—Preprocessing operations, e.g. segment selection; Pattern representation or modelling, e.g. based on linear discriminant analysis [LDA] or principal components; Feature selection or extraction
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
- G10L25/18—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being spectral information of each sub-band
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
- G10L25/21—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being power information
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Acoustics & Sound (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Spectroscopy & Molecular Physics (AREA)
- General Physics & Mathematics (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
- Remote Monitoring And Control Of Power-Distribution Networks (AREA)
Abstract
The invention provides a transformer running state online detection method based on voiceprint recognition, and belongs to the technical field of transformer fault detection. The method comprises the following steps: s101, performing framing processing on the acquired voiceprint signals when the transformer normally operates, and solving a feature vector of each frame of voiceprint signals; s102, fusing all characteristic values in the characteristic vector to obtain a comprehensive evaluation index and the weight of each characteristic value; s103, if the comprehensive evaluation index follows normal distribution, an alarm line is obtained for the comprehensive evaluation index by using a3 sigma rule in statistics; and S104, calculating a corresponding comprehensive evaluation index of the transformer in the unknown operation state according to the weight of each obtained characteristic value, and if the calculated comprehensive evaluation index continuously exceeds an alarm line for multiple times, judging that the transformer in the unknown operation state is abnormal. By adopting the invention, the detection rate of the abnormally operated transformer can be improved, the false detection rate is reduced, and a large number of fault samples are not needed.
Description
Technical Field
The invention relates to the technical field of transformer fault detection, in particular to a transformer running state online detection method based on voiceprint recognition.
Background
In recent years, with the rapid development of extra-high voltage backbone network frames, the strong direct-current power grid structure brings huge pressure to the safe and stable operation of a power grid. The extra-high voltage part equipment is high in technical complexity and still in a quality unstable period, and once a fault occurs, the production and the life are greatly influenced. The traditional detection method mainly comprises the following steps: oil chromatography, ultrasonic detection, vibration detection, etc.; wherein the content of the first and second substances,
the oil chromatographic detection method judges the fault type and the severity degree by detecting the content of the gas component in the oil, and the method has no timeliness because a certain time is provided from the fault occurrence to the change of the gas component in the oil;
the ultrasonic detection method is mainly used for detecting and positioning partial discharge faults at present, has good identification capability only for the partial discharge faults, and has limited detection capability for other fault modes;
the vibration detection method is used for extracting a transformer vibration signal through a sensor tightly attached to the surface of a transformer, and the method is used as a contact detection means, and the detection effect of the transformer on the running state is directly influenced by the installation position and the number of the sensors.
In summary, because the traditional detection method has the limitations of incomplete detection, inconvenient use and the like, a new detection means is required to realize the online detection of the running state of the transformer.
The transformer can generate voiceprint signals with different frequencies in the operation process and radiate outwards through an air medium, and the voiceprint signals mainly comprise voiceprint signals of a transformer body structure, voiceprint signals of a cooling device during working, voiceprint signals generated during working of an on-load tap-changer and the like. Of the voiceprint signals, the dominant transformer body voiceprint signal has a very obvious spectrum characteristic, namely, the 100Hz waveform is taken as a fundamental frequency in normal operation and is mixed with a higher harmonic component. The frequency of the voiceprint signal caused by the operation of the cooling system is typically concentrated in a low frequency region within 100 Hz. And the voiceprint signal generated when the on-load tap changer works is generally a high-frequency component. Therefore, the transformer voiceprint signal contains important information capable of reflecting the operation state of the transformer, and the transformer fault can be effectively detected by analyzing the voiceprint characteristics of the transformer in different operation states. The voiceprint signal can be collected through the acoustic sensor, the acoustic sensor is a non-contact sensor, and compared with a vibration acceleration sensor, the voiceprint signal collection device is more flexible, and the voiceprint signal collection device can be completed without contacting a transformer body.
The existing feature extraction methods mainly include time-Frequency domain features for extracting signals such as feature Frequency, kurtosis and mean values, energy features extraction by adopting wavelet decomposition or wavelet packet decomposition, Mel-Frequency Cepstral Coefficients (MFCCs) feature vectors extraction and the like, and can effectively extract effective information of the transformer vibration and sound signals, but the quantity and selection of feature values are a difficulty. If too many extracted characteristic values exist, the sensitivity of the system is reduced, and the fault detection rate is low; if the extracted characteristic value is too small, the signal characteristics of the vibration sound signal cannot be comprehensively mastered, so that the stability of the system is poor and the false alarm rate is too high. On the other hand, in the feature recognition, the methods adopted by the existing research and patents mainly include Vector Quantization (VQ), Hidden Markov Model (HMM), Support Vector Machine (SVM), expert analysis system, correlation analysis, and the like, and the biggest problems of these methods are: a reasonable classification model can be constructed only on the premise that a large number of fault samples are needed, but the number of fault samples of the transformer is often small, and how to effectively identify faults under the condition of an unbalanced data set is a core key problem to be solved urgently.
Disclosure of Invention
The embodiment of the invention provides a transformer running state online detection method based on voiceprint recognition, through weight analysis, the detection rate of an abnormally running transformer can be improved, the false detection rate is reduced, a large number of fault samples are not needed, and only voiceprint signal samples when the transformer normally runs need to be trained, so that the method is effective for all abnormal states (or fault types), and the application range is greatly improved. The technical scheme is as follows:
in one aspect, a transformer operation state online detection method based on voiceprint recognition is provided, and the method comprises the following steps:
s101, performing framing processing on the acquired voiceprint signals when the transformer normally operates, and solving a feature vector of each frame of voiceprint signals;
s102, fusing all characteristic values in the characteristic vector to obtain a comprehensive evaluation index and the weight of each characteristic value;
s103, judging whether the obtained comprehensive evaluation index is in accordance with normal distribution or not, and if so, solving an alarm line for the comprehensive evaluation index by using a3 sigma rule in statistics;
and S104, calculating a corresponding comprehensive evaluation index of the transformer in the unknown operation state according to the weight of each obtained characteristic value, and if the calculated comprehensive evaluation index continuously exceeds an alarm line for multiple times, judging that the transformer in the unknown operation state is abnormal.
Further, the framing the acquired voiceprint signals of the transformer during normal operation, and solving the feature vector of each frame of voiceprint signals includes:
and performing framing processing on the acquired voiceprint signals of the transformer in normal operation by taking a preset time interval as a frame, and solving 13 characteristic values of the mean value, the variance, the peak value, the peak-to-peak value, the root mean square value, the skewness index, the kurtosis index, the margin index, the pulse index, the peak index, the pass frequency value, the frequency spectrum energy and the main frequency of each frame of voiceprint signals to form a characteristic vector.
Further, the fusing the feature values in the feature vector to obtain the comprehensive evaluation index and the weight of each feature value includes:
a1, carrying out normalization processing on an m multiplied by n dimensional data matrix formed by m frames of voiceprint signals obtained after the framing processing, wherein n represents the number of characteristic values in a characteristic vector;
a2, determining the information entropy e of the jth eigenvalue in the normalized data matrixjAnd according to the obtained information entropy ejDetermining the information utility value d of the jth characteristic valuej；
A3, according to the obtained information utility value djDetermining the weight W of the jth feature valuej；
A4, determining a comprehensive evaluation index based on the weights of all the obtained feature values.
Further, the expression of the normalization process is:
wherein x isijFor the j characteristic value, y, of the i-th frame of voiceprint signal before normalization processingijIs the j characteristic value, x, of the normalized ith frame voiceprint signalj min、xj maxRespectively representing the minimum value and the maximum value of the jth characteristic value in the data matrix.
Further, the information entropy ejExpressed as:
wherein the content of the first and second substances,
denotes yijThe proportion of the sum of jth characteristic value of m frames of voiceprint signals after normalization processing, yijIs the j characteristic value of the normalized ith frame voiceprint signal.
Further, the information utility value djExpressed as:
dj＝1-ej。
further, the weight WjExpressed as:
further, if the normal distribution is obeyed, finding an alarm line for the comprehensive evaluation index by using a3 σ criterion in statistics includes:
if the normal distribution is obeyed, 2 sigma and 3 sigma values are obtained for the comprehensive evaluation index by using a3 sigma rule in statistics, and the 2 sigma and 3 sigma values are respectively used as a primary alarm line and a secondary alarm line.
Further, the method further comprises:
and if the normal distribution is not complied with, performing normal conversion on the comprehensive evaluation index, solving 2 sigma and 3 sigma values of the normally converted comprehensive evaluation index by using a3 sigma rule in statistics, and respectively taking the 2 sigma and 3 sigma values as a primary alarm line and a secondary alarm line.
The technical scheme provided by the embodiment of the invention has the beneficial effects that at least:
in the embodiment of the invention, through weight analysis, more important characteristic values can be highlighted, so that the characteristic values occupy higher proportion, thereby being beneficial to improving the detection rate of the abnormally operated transformer and reducing the false detection rate; and the alarm line is obtained from the comprehensive evaluation index by using the 3 sigma criterion in statistics, so that a large number of fault samples are not needed, and only the voiceprint signal samples in the normal operation of the transformer need to be trained, therefore, the method is effective to all abnormal states (or fault types), and the application range is greatly improved.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a schematic flow chart of a transformer operation state online detection method based on voiceprint recognition according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a transformer operation state online detection system based on voiceprint recognition according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a training and testing process of the transformer operation state online detection method based on voiceprint recognition according to an embodiment of the present invention;
fig. 4 is a schematic diagram of an alarm line of a 220kV outdoor transformer of a certain substation provided by an embodiment of the present invention;
fig. 5 is a schematic diagram of a detection result of a simulated loosening fault of an iron core according to an embodiment of the present invention;
fig. 6 is a schematic diagram of a detection result of a deformation fault of a simulation winding according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
As shown in fig. 1, an embodiment of the present invention provides an online detection method for an operating state of a transformer based on voiceprint recognition, where the method includes:
s101, performing framing processing on the acquired voiceprint signals when the transformer normally operates, and solving a feature vector of each frame of voiceprint signals;
s102, fusing all characteristic values in the characteristic vector to obtain a comprehensive evaluation index and the weight of each characteristic value;
s103, judging whether the obtained comprehensive evaluation index is in accordance with normal distribution or not, and if so, solving an alarm line for the comprehensive evaluation index by using a3 sigma rule in statistics;
and S104, calculating a corresponding comprehensive evaluation index of the transformer in the unknown operation state according to the weight of each obtained characteristic value, and if the calculated comprehensive evaluation index continuously exceeds an alarm line for multiple times, judging that the transformer in the unknown operation state is abnormal.
According to the transformer operation state online detection method based on voiceprint recognition, more important characteristic values can be highlighted through weight analysis, so that the more important characteristic values can occupy higher proportion, the detection rate of the transformer in abnormal operation can be improved, and the false detection rate is reduced; and the alarm line is obtained from the comprehensive evaluation index by using the 3 sigma criterion in statistics, so that a large number of fault samples are not needed, and only the voiceprint signal samples in the normal operation of the transformer need to be trained, therefore, the method is effective to all abnormal states (or fault types), and the application range is greatly improved.
The online detection method for the running state of the transformer based on voiceprint recognition provided by the embodiment of the invention is realized by the online detection system for the running state of the transformer based on voiceprint recognition shown in fig. 2, wherein the system consists of a plurality of acoustic sensors, a multi-channel data acquisition card and a computer, and a data storage and analysis module programmed by labview is arranged in the computer and is used for executing S101-S104.
In the embodiment, the type of the acoustic sensor can be GRAS 40PP CCP Free-field QC Microphone and is used for collecting voiceprint signals of the transformer, so that the running state of the transformer can be effectively detected in a non-contact manner in real time; the sampling frequency of the data acquisition card is 51.2kHz, and the precision is 24 bits. The acoustic sensors are respectively arranged on the periphery of the transformer, connected to the input end of the data acquisition card and used for multi-channel synchronous acquisition. The output end of the data acquisition card is connected to the network port end of the computer for data display, storage and analysis.
In a specific implementation manner of the foregoing method for online detecting an operating state of a transformer based on voiceprint recognition, further, the framing the acquired voiceprint signals during normal operation of the transformer, and obtaining a feature vector of each frame of voiceprint signals includes:
the method comprises the steps of performing framing processing on collected voiceprint signals when a transformer normally runs by taking a preset time interval (for example, 1s) as one frame, and obtaining 13 feature values of a mean value, a variance, a peak value, a peak-peak value, a root mean square value, a skewness index, a kurtosis index, a margin index, a pulse index, a peak index, a pass frequency value, spectrum energy and a main frequency of each frame of voiceprint signals to form a feature vector.
In this embodiment, S101 to S103 execute the training process in fig. 3, and first acquire a large number of voiceprint signals during normal operation of the transformer as training samples, and then perform framing processing on training set data formed by the training samples by taking 1S as a frame to obtain a feature vector formed by the 13 feature values, so that the framing interval can be shortened, the data size can be increased, and richer voiceprint information can be acquired without losing spectral features. Assuming that m seconds of transformer voiceprint signals are acquired, m frames of voiceprint signals are shared, and the 13 eigenvalues are calculated for each frame, so that an m × n data matrix can be obtained, where n represents the number of eigenvalues in the eigenvector, and n is 13.
In this embodiment, the importance of different characteristic values is different, which means that the contribution values of different characteristic values to the evaluation of the transformer operating state are different, so that the importance of different characteristic values needs to be evaluated, and the evaluation criterion may be measured by a weight, where a larger weight indicates that the contribution value of the characteristic value to the evaluation of the transformer operating state is larger, and vice versa. In the present application, the entropy weight method may be used to analyze each eigenvalue (i.e., 13 eigenvalues) in the eigenvector to obtain weights of the 13 eigenvalues, and finally, according to the obtained weights of each eigenvalue, all eigenvalues are fused into a comprehensive evaluation index (MVF), which specifically includes the following steps:
a1, carrying out normalization processing on an m multiplied by n dimensional data matrix formed by m frames of voiceprint signals obtained after the framing processing, wherein n represents the number of characteristic values in a characteristic vector;
in this embodiment, the expression of the normalization process is:
wherein x isijFor the j characteristic value, y, of the i-th frame of voiceprint signal before normalization processingijIs the j characteristic value, x, of the normalized ith frame voiceprint signalj min、xj maxRespectively representing the minimum value and the maximum value of the j-th characteristic value in the data matrix, wherein the maximum value of i is m, and the maximum value of j is 13.
A2, determining the information entropy e of the jth eigenvalue in the normalized data matrixjAnd according to the obtained information entropy ejDetermining the information utility value d of the jth characteristic valuej；
In this embodiment, the information entropy value ejExpressed as:
wherein the content of the first and second substances,
denotes yijThe proportion of the sum of jth characteristic value of m frames of voiceprint signals after normalization processing, yijIs the j characteristic value of the normalized ith frame voiceprint signal.
In this embodiment, the information utility value djExpressed as:
dj＝1-ej。
a3, according to the obtained information utility value djDetermining the weight W of the jth feature valuej；
In this embodiment, the weight WjExpressed as:
from weight WjAs can be seen from the expression of (c), the weight WjThe cumulative sum of (c) is 1.
A4, determining a comprehensive evaluation index (MVF) according to the weight of all the obtained characteristic values, wherein the MVF is expressed as:
in a specific embodiment of the above online detection method for the operating state of the transformer based on voiceprint recognition, further, if the normal distribution is obeyed, finding the alarm line for the comprehensive evaluation index by using a3 σ criterion in statistics includes:
if the normal distribution is obeyed, 2 sigma and 3 sigma values are obtained for the comprehensive evaluation index by using a3 sigma rule in statistics, and the 2 sigma and 3 sigma values are respectively used as a primary alarm line and a secondary alarm line.
In this embodiment, a K-S inspection method may be used to determine whether the obtained comprehensive evaluation index (MVF) complies with normal distribution, and if the obtained comprehensive evaluation index (MVF) complies with normal distribution, a3 σ criterion in statistics is used to obtain 2 σ and 3 σ values for the comprehensive evaluation index, and the 2 σ and 3 σ values are respectively used as a primary alarm line and a secondary alarm line;
and if the normal distribution is not complied with, performing normal conversion on the comprehensive evaluation index, solving 2 sigma and 3 sigma values of the normally converted comprehensive evaluation index by using a3 sigma rule in statistics, and respectively taking the 2 sigma and 3 sigma values as a primary alarm line and a secondary alarm line.
The alarm line of the 220kV outdoor transformer of a certain transformer substation obtained by the method is shown in fig. 4, wherein high load and low load can be judged according to a load diagram of the transformer.
In this embodiment, through S101 to S103, an alarm line has been established for the normal operation state of the transformer. For a transformer of unknown operating state, a test procedure as in fig. 3 is performed: acquiring a voiceprint signal of a transformer in an unknown running state as a test sample, performing frame division processing on test set data formed by the test sample by taking 1s as a frame, obtaining 13 characteristic values of the test set data, fusing the 13 characteristic values into 1 comprehensive evaluation index according to the weight of a training set, entering an observation state if the calculated comprehensive evaluation index exceeds a primary alarm line for multiple times, and judging that the transformer in the unknown running state is abnormal if the calculated comprehensive evaluation index exceeds a secondary alarm line for multiple times continuously.
In order to verify the effectiveness of the transformer running state online detection method and system based on voiceprint recognition provided by the embodiment of the invention, the method and system are applied to iron core loosening fault detection and winding deformation fault detection, the simulated iron core loosening fault detection result is shown in fig. 5, and the simulated winding deformation fault detection result is shown in fig. 6.
In summary, the transformer operation state online detection method and system based on voiceprint recognition provided by the embodiment of the invention have the following effective effects:
1) the detection method provided by the invention has real-time performance, and can feed back abnormal information in time when the transformer is abnormal, thereby avoiding extra loss and serious disaster accidents;
2) the detection method provided by the invention can effectively detect the running state of the transformer in real time under the condition of not contacting with the transformer body, can replace manual inspection to a certain extent, and reduces human errors while improving the working efficiency;
3) the detection method provided by the invention can highlight more important characteristic values through weight analysis, so that the characteristic values occupy higher proportion, thereby improving the detection rate of the abnormally operated transformer, reducing the false detection rate and improving the stability of the system;
4) the detection method based on the 3 sigma criterion does not need a large number of fault samples, and only needs to train the voiceprint signal samples when the transformer normally operates, so that the method is effective to all abnormal states (or fault types), and the application range of the method is greatly improved.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (2)
1. A transformer running state online detection method based on voiceprint recognition is characterized by comprising the following steps:
s101, performing framing processing on the acquired voiceprint signals when the transformer normally operates, and solving a feature vector of each frame of voiceprint signals;
s102, fusing all characteristic values in the characteristic vector to obtain a comprehensive evaluation index and the weight of each characteristic value;
s103, judging whether the obtained comprehensive evaluation index is in accordance with normal distribution or not, and if so, solving an alarm line for the comprehensive evaluation index by using a3 sigma rule in statistics;
s104, for the transformer in the unknown operation state, calculating a corresponding comprehensive evaluation index according to the weight of each obtained characteristic value, and if the calculated comprehensive evaluation index continuously exceeds an alarm line for multiple times, judging that the transformer in the unknown operation state is abnormal;
the framing processing of the acquired voiceprint signals during normal operation of the transformer and the calculation of the feature vector of each frame of voiceprint signals comprise:
performing framing processing on the collected voiceprint signals of the transformer in normal operation by taking a preset time interval as a frame, and solving 13 characteristic values of a mean value, a variance, a peak value, a peak-to-peak value, a root mean square value, a skewness index, a kurtosis index, a margin index, a pulse index, a peak index, a pass frequency value, frequency spectrum energy and a main frequency of each frame of voiceprint signals to form a characteristic vector;
the method for obtaining the comprehensive evaluation index and the weight of each characteristic value by fusing each characteristic value in the characteristic vector comprises the following steps:
a1, carrying out normalization processing on an m multiplied by n dimensional data matrix formed by m frames of voiceprint signals obtained after the framing processing, wherein n represents the number of characteristic values in a characteristic vector;
a2, determining the information entropy e of the jth eigenvalue in the normalized data matrixjAnd according to the obtained information entropy ejDetermining the information utility value d of the jth characteristic valuej；
A3, according to the obtained information utility value djDetermining the weight W of the jth feature valuej；
A4, determining a comprehensive evaluation index according to the weight of all the obtained characteristic values;
wherein the information entropy ejExpressed as:
wherein the content of the first and second substances,
denotes yijThe proportion of the sum of jth characteristic value of m frames of voiceprint signals after normalization processing, yijThe j characteristic value of the i frame of the voiceprint signal after normalization processing is obtained;
wherein the information utility value djExpressed as:
dj＝1-ej；
wherein the weight WjExpressed as:
wherein the comprehensive evaluation index is expressed as:
wherein, MVF represents a comprehensive evaluation index;
wherein, if the normal distribution is obeyed, the step of obtaining an alarm line for the comprehensive evaluation index by using a3 sigma rule in statistics comprises the following steps:
if the normal distribution is obeyed, 2 sigma and 3 sigma values are obtained for the comprehensive evaluation index by using a3 sigma rule in statistics, and the 2 sigma and 3 sigma values are respectively used as a primary alarm line and a secondary alarm line;
wherein the method further comprises:
and if the normal distribution is not complied with, performing normal conversion on the comprehensive evaluation index, solving 2 sigma and 3 sigma values of the normally converted comprehensive evaluation index by using a3 sigma rule in statistics, and respectively taking the 2 sigma and 3 sigma values as a primary alarm line and a secondary alarm line.
2. The online detection method for the running state of the transformer based on the voiceprint recognition is characterized in that the expression of normalization processing is as follows:
wherein x isijFor the j characteristic value, y, of the i-th frame of voiceprint signal before normalization processingijIs the j characteristic value, x, of the normalized ith frame voiceprint signalj min、xj maxRespectively representing the minimum value and the maximum value of the jth characteristic value in the data matrix.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010930602.2A CN112201260B (en) | 2020-09-07 | 2020-09-07 | Transformer running state online detection method based on voiceprint recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010930602.2A CN112201260B (en) | 2020-09-07 | 2020-09-07 | Transformer running state online detection method based on voiceprint recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112201260A CN112201260A (en) | 2021-01-08 |
| CN112201260B true CN112201260B (en) | 2021-10-12 |
Family
ID=74005378
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010930602.2A Active CN112201260B (en) | 2020-09-07 | 2020-09-07 | Transformer running state online detection method based on voiceprint recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112201260B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112946442B (en) * | 2021-04-12 | 2022-06-28 | 厦门理工学院 | Switch cabinet partial discharge detection method, terminal equipment and storage medium |
| CN113314142B (en) * | 2021-05-26 | 2022-04-01 | 山东大学 | Engine fault detection method, system, storage medium and equipment |
| CN113436645A (en) * | 2021-06-09 | 2021-09-24 | 杭州电子科技大学 | Electromechanical system fault on-line monitoring acoustic processing method under complex noise environment |
| CN113674447A (en) * | 2021-07-09 | 2021-11-19 | 深圳市慧友安电子技术有限公司 | System and method for detecting voiceprint fault of industrial equipment |
| CN113674764A (en) * | 2021-08-20 | 2021-11-19 | 广东外语外贸大学 | Interpretation evaluation method, system and equipment based on bidirectional cyclic neural network |
| CN113985156A (en) * | 2021-09-07 | 2022-01-28 | 绍兴电力局柯桥供电分局 | Intelligent fault identification method based on transformer voiceprint big data |
| CN114217149A (en) * | 2021-11-14 | 2022-03-22 | 国网辽宁省电力有限公司葫芦岛供电公司 | Transformer acoustic fingerprint uninterrupted power detection and state early warning method |
| CN114937462A (en) * | 2022-05-17 | 2022-08-23 | 国网黑龙江省电力有限公司佳木斯供电公司 | High-voltage circuit breaker fault detection method based on voiceprint intelligent diagnosis |
| CN115358350A (en) * | 2022-10-19 | 2022-11-18 | 深圳市乐闵科技有限公司 | Abnormity alarm method for paper diaper production equipment |
| CN115993503B (en) * | 2023-03-22 | 2023-06-06 | 广东电网有限责任公司东莞供电局 | Operation detection method, device and equipment of transformer and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102411106A (en) * | 2011-11-18 | 2012-04-11 | 广东电网公司广州供电局 | Fault monitoring method and device for power transformer |
| CN104793605A (en) * | 2015-04-10 | 2015-07-22 | 北京金控自动化技术有限公司 | Method for judging equipment faults by means of normal distribution |
| CN110398647A (en) * | 2019-06-26 | 2019-11-01 | 深圳供电局有限公司 | Transformer's Condition Monitoring method |
| CN110490071A (en) * | 2019-07-12 | 2019-11-22 | 国网上海市电力公司 | A kind of substation's Abstraction of Sound Signal Characteristics based on MFCC |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102270271B (en) * | 2011-05-03 | 2014-03-19 | 北京中瑞泰科技有限公司 | Equipment failure early warning and optimizing method and system based on similarity curve |
-
2020
- 2020-09-07 CN CN202010930602.2A patent/CN112201260B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102411106A (en) * | 2011-11-18 | 2012-04-11 | 广东电网公司广州供电局 | Fault monitoring method and device for power transformer |
| CN104793605A (en) * | 2015-04-10 | 2015-07-22 | 北京金控自动化技术有限公司 | Method for judging equipment faults by means of normal distribution |
| CN110398647A (en) * | 2019-06-26 | 2019-11-01 | 深圳供电局有限公司 | Transformer's Condition Monitoring method |
| CN110490071A (en) * | 2019-07-12 | 2019-11-22 | 国网上海市电力公司 | A kind of substation's Abstraction of Sound Signal Characteristics based on MFCC |
Non-Patent Citations (1)
| Title |
|---|
| 基于声音特征的变电设备故障分类与定位方法研究;侯增起;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20190115;C042-2021 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112201260A (en) | 2021-01-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112201260B (en) | Transformer running state online detection method based on voiceprint recognition | |
| CN111626153B (en) | Integrated learning-based partial discharge fault state identification method | |
| CN108169639B (en) | Method for identifying switch cabinet fault based on parallel long-time and short-time memory neural network | |
| CN102426835B (en) | Method for identifying local discharge signals of switchboard based on support vector machine model | |
| CN106934421B (en) | Power transformer fault detection method based on 2DPCA and SVM | |
| CN109949823B (en) | DWPT-MFCC and GMM-based in-vehicle abnormal sound identification method | |
| CN110458248B (en) | Transformer abnormal state detection method based on multi-point vibration signals | |
| CN113124929A (en) | Transformer substation multi-parameter signal acquisition comprehensive analysis system and method | |
| CN110221137B (en) | Distribution transformer abnormal state detection method based on vibration-sound correlation | |
| CN112147444B (en) | Power transformer working state monitoring method and system | |
| CN111948487A (en) | High-voltage power equipment fault diagnosis method and system based on artificial intelligence | |
| CN104156568A (en) | Transformer fault diagnosis method on basis of weighted gray correlation and fuzzy clustering | |
| CN105203936A (en) | Method for determining power cable partial discharge defect type based on spectral analysis | |
| WO2023245990A1 (en) | Classification and abnormality detection method for electrical devices of power plant on basis of multi-information fusion | |
| CN110488675A (en) | A kind of substation's Abstraction of Sound Signal Characteristics based on dynamic time warpping algorithm | |
| CN114217149A (en) | Transformer acoustic fingerprint uninterrupted power detection and state early warning method | |
| CN115839763A (en) | Transformer vibration noise monitoring system | |
| CN116778964A (en) | Power transformation equipment fault monitoring system and method based on voiceprint recognition | |
| CN113345399A (en) | Method for monitoring sound of machine equipment in strong noise environment | |
| CN115376526A (en) | Power equipment fault detection method and system based on voiceprint recognition | |
| CN117169639B (en) | Product detection method and system for power adapter production | |
| CN117607598A (en) | Transformer fault detection method and system based on voiceprint characteristics | |
| CN113362856A (en) | Sound fault detection method and device applied to power Internet of things | |
| CN115410599A (en) | Method and system for identifying running state of transformer based on voiceprint perception | |
| Yang et al. | Transient fault diagnosis of track circuit based on MFCC-DTW |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |