CN116230013A - Transformer fault voiceprint detection method based on x-vector - Google Patents
Transformer fault voiceprint detection method based on x-vector Download PDFInfo
- Publication number
- CN116230013A CN116230013A CN202310184476.4A CN202310184476A CN116230013A CN 116230013 A CN116230013 A CN 116230013A CN 202310184476 A CN202310184476 A CN 202310184476A CN 116230013 A CN116230013 A CN 116230013A
- Authority
- CN
- China
- Prior art keywords
- voiceprint
- vector
- transformer
- data
- fault
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 16
- 238000000034 method Methods 0.000 claims abstract description 19
- 238000013528 artificial neural network Methods 0.000 claims abstract description 11
- 230000005236 sound signal Effects 0.000 claims abstract description 11
- 230000003595 spectral effect Effects 0.000 claims abstract description 7
- 238000012549 training Methods 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 230000009467 reduction Effects 0.000 claims abstract description 5
- 238000001228 spectrum Methods 0.000 claims description 22
- 238000004458 analytical method Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 7
- 238000011176 pooling Methods 0.000 claims description 6
- 238000009432 framing Methods 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 abstract description 5
- 238000005070 sampling Methods 0.000 description 4
- 238000012423 maintenance Methods 0.000 description 3
- 230000035945 sensitivity Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 241000764238 Isis Species 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012360 testing 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
- 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
-
- 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
- G10L17/04—Training, enrolment or model building
-
- 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
- G10L17/18—Artificial neural networks; Connectionist approaches
-
- 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
- G10L25/27—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique
- G10L25/30—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique using neural networks
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
- Y04S10/52—Outage or fault management, e.g. fault detection or location
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The method comprises the steps of selecting a proper acquisition device to acquire data, simply classifying the acquired data according to fault types, and carrying out noise reduction by using spectral subtraction; secondly, preprocessing fault data to obtain a feature vector of a sound signal; then, building an x-vector network framework, and training voiceprint features extracted by preprocessing as the input of a neural network; thirdly, inputting the characteristics of the voiceprint data to be identified into a model, and setting a threshold value to obtain an identification result of the fault state of the transformer according to similarity matching; and finally, designing a transformer voiceprint monitoring platform and displaying the result of transformer voiceprint recognition in real time. The invention provides an x-vector-based transformer fault voiceprint detection method capable of effectively reducing noise interference and improving generalization capability of an identification model.
Description
Technical Field
The invention relates to the field of transformer fault voiceprint recognition and detection, in particular to a transformer fault voiceprint detection method based on an x-vector.
Background
In a power grid, a power transformer is an important energy transmission and conversion device, the number of the power transformers is very large, tens of millions of transformers are expected in China, and millions of transformers need to be updated and maintained every year, and the safe, reliable and stable operation of the transformers is the fundamental of the safe, reliable and stable operation of a power supply network, so that the accurate evaluation of the operation state of the transformers is significant in effectively maintaining the operation stability of the power transformers and improving the maintenance level. Since the mechanical vibration produces sound and has certain regularity when the power transformer works, a method for judging the running state of the transformer according to the voiceprint of the transformer is provided. Along with the rapid development of national economy in China in recent years, higher requirements are put forward on the quality and quantity of power supply and distribution networks. The traditional transformer fault detection method combines manual data sampling and online parameter sampling, and maintenance personnel judge the running state of the transformer according to the sampling data. However, since the manual judgment is generally only combined with the sampling data at the current moment, predictability is lacking, faults cannot be effectively predicted, and immeasurable losses can be caused. Therefore, the introduction of computer on-line detection and auxiliary analysis is particularly important.
With the continuous iteration of the transformer voiceprint recognition algorithm, besides improving a trained model, the accuracy of voiceprint recognition is also in close and inseparable connection with the acquired quality. Noise generated by the external environment in the actual offline acquisition process cannot be effectively removed, so that the noise has more or less influence on subsequent voiceprint recognition judgment. Secondly, because of daily maintenance, the frequency of the transformer faults is flexible, so the difficulty of obtaining the voiceprint data of the transformer faults is greatly improved. The small amount of data can result in a model with insufficient numbers and no way to find the features therein, in which case fitting the data can result in a test error that is high although the training results are not very large, with a gap from being put into use, thus solving such problems.
Disclosure of Invention
In order to solve the problems of noise pollution, insufficient fault data set and the like of the existing transformer voiceprint fault detection method, the invention provides the x-vector-based transformer fault voiceprint detection method capable of effectively reducing noise interference and improving the generalization capability of an identification model.
The technical scheme adopted for solving the technical problems is as follows:
an x-vector based transformer fault detection method, the method comprising the steps of:
1) When voiceprint detection is carried out, a microphone is required to collect sound;
2) Noise reduction is carried out on the voiceprint signal by using spectral subtraction, and the noise power spectrum is subtracted from the noise power spectrum to obtain a denoised sound signal power spectrum;
3) Preprocessing input voiceprint data to obtain an abstract voiceprint feature;
4) And building a neural network structure of an x-vector, wherein the x-vector comprises a TDNN layer with a multi-layer frame level, a statistics pooling layer, a full-connection layer with two sentence levels and a softmax, and the loss function is cross entropy. The x-vector can accept any length of input and combine the frame-level features into the whole sentence features thanks to the statistical pooling layer in the network;
5) Training a data set to learn a recognition model by using the constructed neural network frame;
6) And inputting the feature vector of the data to be tested into the trained model to obtain an output feature of voiceprint data, matching the output feature with a feature template of the current fault type, and judging the state type corresponding to the transformer according to the similarity setting threshold.
Further, the procedure of the step 3) is as follows:
3.1 Pre-emphasis of the sound signal, i.e. a emphasis process of the high frequency signal: the value of each signal minus the value of the next signal multiplied by a coefficient α is given by: y [ n ] = x [ n ] - αx [ n-1];
3.2 Data splicing is carried out on voiceprint data with time smaller than a preset threshold value, and the data is effectively utilized;
3.3 Framing and windowing the signals, performing fast Fourier transform on each short-time analysis window to obtain a corresponding frequency spectrum, and passing the frequency spectrum through a Mel filter to obtain a Mel frequency spectrum; carrying out cepstrum analysis on the Mel frequency spectrum, taking logarithm, carrying out DCT discrete cosine transform, taking the 2 nd to 13 th coefficients after DCT as MFCC coefficients, and obtaining Mel frequency cepstrum coefficient MFCC, wherein the MFCC is the characteristic of the frame of voice;
3.4 After the voiceprint features of the transformer are extracted, modeling the target working condition, comparing and judging the noise signal to be detected, and identifying the fault working condition type; the method comprises the steps of intelligently identifying 3 fault working condition types including DC magnetic bias, accessory looseness and partial discharge of a transformer, and sending the extracted characteristics into a neural network of an x-vector.
Preferably, in step 1), the microphone is a capacitive sensor, and sound is collected by the capacitive microphone and transmitted to the computer.
The technical conception of the invention is as follows: firstly, selecting a proper acquisition device to acquire data, simply classifying the acquired data according to fault types, and carrying out noise reduction by using spectral subtraction; secondly, preprocessing fault data to obtain a feature vector of a sound signal; then, building an x-vector network framework, and training voiceprint features extracted by preprocessing as the input of a neural network; thirdly, inputting the characteristics of the voiceprint data to be identified into a model, and setting a threshold value to obtain an identification result of the fault state of the transformer according to similarity matching; and finally, designing a transformer voiceprint monitoring platform and displaying the result of transformer voiceprint recognition in real time.
The beneficial effects of the invention are as follows: firstly, performance indexes such as sensitivity and reliability are integrated, a voiceprint acquisition device is selected, and acquired data is simply processed. And secondly, the influence of environmental noise on the target voiceprint is reduced by using spectral subtraction, and the recognition degree of the fault voiceprint of the transformer is improved. Finally, designing a transformer voiceprint monitoring platform to realize real-time display and statistics of voiceprint recognition results;
drawings
Fig. 1 is a schematic diagram of basic steps of a transformer fault voiceprint detection method.
Fig. 2 shows a network structure of an x-vector.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
Referring to fig. 1 and 2, an x-vector based transformer fault voiceprint detection method, the prediction method includes the following steps:
1) When voiceprint detection is carried out, a microphone is required to collect sound, so that the selection of the microphone is particularly important, through analysis and consideration of aspects such as microphone sensitivity, reliability and the like, the microphone selected by the user needs to be small in size and convenient to install and carry, and the normal operation of the transformer cannot be influenced when voiceprint data are collected close to the transformer, so that the capacitive sensor is selected. The acquisition of sound is performed by a condenser microphone, which has no coil or magnet, and voltage variation is generated by the distance between two separators of the condenser, and the condenser sensor is commonly used for high-quality recording due to high sensitivity. The microphone incorporates a sound sensitive capacitive electret microphone. The sound wave vibrates the electret film in the microphone, and causes a change in capacitance, thereby generating a minute voltage that changes in response thereto. This voltage is then converted to a voltage of 0-5V, is received by the data collector via a/D conversion, and is transmitted to the computer;
the collected data is simply processed, and the collected data comprises three fault samples of direct current magnetic bias, fan aging and partial discharge and normal sound samples. Modifying the format of the sound data, and converting the sound data into a wav format;
2) In the process of collecting sound data, various interference noise often exists, and the processing and fault diagnosis of the body sound of the transformer are puzzled. The noise reduction method for the transformer voiceprint recognition comprises a wavelet threshold denoising method, a FastICA-based denoising method and the like, wherein the traditional spectral subtraction is adopted. The principle of spectral subtraction is as follows: let s (t) be the denoised sound signal, d (t) be the noise signal, and y (t) be the original sound signal: y (t) =s (t) +d (t)
The fourier transforms of Y (ω), S (ω), D (ω) are Y (t), S (t), D (t), respectively, can be derived:
|Y(ω)| 2 ＝|S(ω)| 2 +|D(ω)| 2
P y (ω)，P s (ω)，P d and (ω) is the power spectrum of y (t), s (t), d (t), respectively:
P y (ω)＝P s (ω)+P d (ω)
namely, the noise power spectrum is subtracted from the noise power spectrum to obtain the denoised sound signal power spectrum.
3) The fault data is preprocessed, and the process of obtaining the feature vector of the sound signal is as follows:
3.1 Pre-emphasis of the sound signal, i.e. a emphasis process of the high frequency signal: the value of subtracting a larger coefficient from each signal, such as 0.95 or 0.97, is multiplied by the following equation: y [ n ] = x [ n ] - αx [ n-1];
3.2 Data splicing is carried out on voice print data with shorter time, so that the influence caused by small data quantity is reduced, and the data is effectively utilized;
3.3 Framing and windowing the signals, performing fast Fourier transform on each short-time analysis window to obtain a corresponding frequency spectrum, and passing the frequency spectrum through a Mel filter to obtain a Mel frequency spectrum; carrying out cepstrum analysis on the Mel frequency spectrum, taking logarithm, carrying out DCT discrete cosine transform, taking 50 coefficients after DCT as MFCC coefficients, and obtaining Mel frequency cepstrum coefficient MFCC which is the characteristic of the frame of voice;
3.4 After the voiceprint features of the transformer are extracted, modeling the target working condition is needed, comparing and judging the noise signal to be detected, and identifying the fault working condition type; aiming at DC magnetic bias of the transformer and loosening accessories, carrying out intelligent identification on the partial discharge total 3 fault working condition types, and sending the extracted characteristics into a neural network of an x-vector;
4) And building a neural network structure of an x-vector, wherein the x-vector comprises a TDNN layer with a multi-layer frame level, a statistics pooling layer, a full-connection layer with two sentence levels and a softmax, and the loss function is cross entropy. The x-vector can accept any length of input and combine the frame-level features into the whole sentence features thanks to the statistical pooling layer in the network;
5) Inputting the training set into a neural network for training to obtain a learning model;
6) And inputting the feature vector of the data to be tested into the trained model to obtain an output feature of voiceprint data, matching the output feature with a feature template of the current fault type, and judging the state type corresponding to the transformer according to the similarity setting threshold.
The embodiment designs a transformer voiceprint monitoring platform, and displays the result of transformer voiceprint recognition in real time, wherein the result comprises modules of overall state, real-time waveform, anomaly statistics, history alarm, equipment management and the like. Firstly, a general framework of each module is designed, and detailed drawing of pages is carried out by using tools such as jQuery, bootStrap, echarts and the like. And acquiring data by using Ajax, acquiring the related state information of the database transformer, and rendering the related state information into the page. Finally, the web interface can be deployed on line by utilizing the isis service, so that the voiceprint monitoring platform can be accessed through the IP address of the front-end server under the same local area network;
the foregoing description of the preferred embodiment of the present invention is not intended to limit the invention, but rather to make various modifications and improvements without departing from the scope of the invention as set forth in the essential teachings of the invention.
Claims (3)
1. An x-vector based transformer fault detection method, comprising the steps of:
1) When voiceprint detection is carried out, a microphone is required to collect sound;
2) Noise reduction is carried out on the voiceprint signal by using spectral subtraction, and the noise power spectrum is subtracted from the noise power spectrum to obtain a denoised sound signal power spectrum;
3) Preprocessing input voiceprint data to obtain an abstract voiceprint feature;
4) And building a neural network structure of an x-vector, wherein the x-vector comprises a TDNN layer with a multi-layer frame level, a statistics pooling layer, a full-connection layer with two sentence levels and a softmax, and the loss function is cross entropy. The x-vector can accept any length of input and combine the frame-level features into the whole sentence features thanks to the statistical pooling layer in the network;
5) Training a data set to learn a recognition model by using the constructed neural network frame;
6) And inputting the feature vector of the data to be tested into the trained model to obtain an output feature of voiceprint data, matching the output feature with a feature template of the current fault type, and judging the state type corresponding to the transformer according to the similarity setting threshold.
2. The method for detecting the fault voiceprint of the transformer based on the x-vector as set forth in claim 1, wherein the process of the step 3) is as follows:
3.1 Pre-emphasis of the sound signal, i.e. a emphasis process of the high frequency signal: the value of each signal minus the value of the next signal multiplied by a coefficient α is given by: y [ n ] = x [ n ] - αx [ n-1];
3.2 Data splicing is carried out on voiceprint data with time smaller than a preset threshold value, and the data is effectively utilized;
3.3 Framing and windowing the signals, performing fast Fourier transform on each short-time analysis window to obtain a corresponding frequency spectrum, and passing the frequency spectrum through a Mel filter to obtain a Mel frequency spectrum; carrying out cepstrum analysis on the Mel frequency spectrum, taking logarithm, carrying out DCT discrete cosine transform, taking the 2 nd to 13 th coefficients after DCT as MFCC coefficients, and obtaining Mel frequency cepstrum coefficient MFCC, wherein the MFCC is the characteristic of the frame of voice;
3.4 After the voiceprint features of the transformer are extracted, modeling the target working condition, comparing and judging the noise signal to be detected, and identifying the fault working condition type; the method comprises the steps of intelligently identifying 3 fault working condition types including DC magnetic bias, accessory looseness and partial discharge of a transformer, and sending the extracted characteristics into a neural network of an x-vector.
3. The method for detecting fault voiceprint of transformer based on x-vector as claimed in claim 1 or 2, wherein in step 1), the microphone is a capacitive sensor, and sound is collected by the capacitive microphone and transmitted to the computer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310184476.4A CN116230013A (en) | 2023-02-27 | 2023-02-27 | Transformer fault voiceprint detection method based on x-vector |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310184476.4A CN116230013A (en) | 2023-02-27 | 2023-02-27 | Transformer fault voiceprint detection method based on x-vector |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116230013A true CN116230013A (en) | 2023-06-06 |
Family
ID=86576504
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310184476.4A Pending CN116230013A (en) | 2023-02-27 | 2023-02-27 | Transformer fault voiceprint detection method based on x-vector |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116230013A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117470976A (en) * | 2023-12-28 | 2024-01-30 | 烟台宇控软件有限公司 | Transmission line defect detection method and system based on voiceprint features |
| CN117894317A (en) * | 2024-03-14 | 2024-04-16 | 沈阳智帮电气设备有限公司 | Box-type transformer on-line monitoring method and system based on voiceprint analysis |
| CN117894317B (en) * | 2024-03-14 | 2024-05-24 | 沈阳智帮电气设备有限公司 | Box-type transformer on-line monitoring method and system based on voiceprint analysis |
-
2023
- 2023-02-27 CN CN202310184476.4A patent/CN116230013A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117470976A (en) * | 2023-12-28 | 2024-01-30 | 烟台宇控软件有限公司 | Transmission line defect detection method and system based on voiceprint features |
| CN117470976B (en) * | 2023-12-28 | 2024-03-26 | 烟台宇控软件有限公司 | Transmission line defect detection method and system based on voiceprint features |
| CN117894317A (en) * | 2024-03-14 | 2024-04-16 | 沈阳智帮电气设备有限公司 | Box-type transformer on-line monitoring method and system based on voiceprint analysis |
| CN117894317B (en) * | 2024-03-14 | 2024-05-24 | 沈阳智帮电气设备有限公司 | Box-type transformer on-line monitoring method and system based on voiceprint analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111325095B (en) | Intelligent detection method and system for equipment health state based on acoustic wave signals | |
| CN109034046B (en) | Method for automatically identifying foreign matters in electric energy meter based on acoustic detection | |
| CN113707176B (en) | Transformer fault detection method based on acoustic signal and deep learning technology | |
| CN111912519B (en) | Transformer fault diagnosis method and device based on voiceprint frequency spectrum separation | |
| CN115932659A (en) | Transformer fault detection method based on voiceprint characteristics | |
| CN111948487A (en) | High-voltage power equipment fault diagnosis method and system based on artificial intelligence | |
| WO2022088643A1 (en) | Fault diagnosis method and apparatus for buried transformer substation, and electronic device | |
| CN116230013A (en) | Transformer fault voiceprint detection method based on x-vector | |
| CN116778964A (en) | Power transformation equipment fault monitoring system and method based on voiceprint recognition | |
| CN115376526A (en) | Power equipment fault detection method and system based on voiceprint recognition | |
| CN113345399A (en) | Method for monitoring sound of machine equipment in strong noise environment | |
| CN116741148A (en) | Voice recognition system based on digital twinning | |
| CN114722895A (en) | Photovoltaic array fault diagnosis method based on non-invasive state detection | |
| CN114487733A (en) | Partial discharge detection method based on voiceprint | |
| CN115618205A (en) | Portable voiceprint fault detection system and method | |
| Rahman et al. | Dynamic time warping assisted svm classifier for bangla speech recognition | |
| Qi et al. | Feature classification method of frequency cepstrum coefficient based on weighted extreme gradient boosting | |
| Sun et al. | Jointed Task of Multi-scale CNN Based Power Transformer Fault Diagnosis with Vibration and Sound Signals | |
| CN114157023B (en) | Distribution transformer early warning information acquisition method | |
| CN112735477B (en) | Voice emotion analysis method and device | |
| Li et al. | A Robust Classification Method for Power Transformer Status Recognition Based on Sound Signals | |
| CN111091816A (en) | Data processing system and method based on voice evaluation | |
| CN115662460B (en) | Abnormal sound extraction method for industrial equipment | |
| CN116717461B (en) | Intelligent monitoring method and system for operating state of vacuum pump | |
| CN117524250A (en) | Hydroelectric generating set abnormal sound detection method based on self-supervision and weak supervision |
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 |