CN115358110A - Transformer fault detection system based on acoustic sensor array - Google Patents
Transformer fault detection system based on acoustic sensor array Download PDFInfo
- Publication number
- CN115358110A CN115358110A CN202210876670.4A CN202210876670A CN115358110A CN 115358110 A CN115358110 A CN 115358110A CN 202210876670 A CN202210876670 A CN 202210876670A CN 115358110 A CN115358110 A CN 115358110A
- Authority
- CN
- China
- Prior art keywords
- transformer
- vibration
- acoustic
- fault
- noise
- 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 26
- 238000004804 winding Methods 0.000 claims abstract description 58
- 238000004458 analytical method Methods 0.000 claims abstract description 24
- 230000005855 radiation Effects 0.000 claims abstract description 23
- 238000003745 diagnosis Methods 0.000 claims abstract description 21
- 230000002159 abnormal effect Effects 0.000 claims abstract description 17
- 238000012544 monitoring process Methods 0.000 claims abstract description 17
- 230000008878 coupling Effects 0.000 claims abstract description 14
- 238000010168 coupling process Methods 0.000 claims abstract description 14
- 238000005859 coupling reaction Methods 0.000 claims abstract description 14
- 238000005516 engineering process Methods 0.000 claims abstract description 14
- 238000004088 simulation Methods 0.000 claims abstract description 14
- 238000012545 processing Methods 0.000 claims abstract description 11
- 230000005236 sound signal Effects 0.000 claims abstract description 10
- 238000011156 evaluation Methods 0.000 claims abstract description 8
- 238000000034 method Methods 0.000 claims description 36
- 230000005540 biological transmission Effects 0.000 claims description 22
- 238000013507 mapping Methods 0.000 claims description 13
- 238000012360 testing method Methods 0.000 claims description 10
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical group [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 claims description 9
- 230000008859 change Effects 0.000 claims description 8
- 230000006870 function Effects 0.000 claims description 7
- 238000013527 convolutional neural network Methods 0.000 claims description 6
- 238000009826 distribution Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 6
- 238000003384 imaging method Methods 0.000 claims description 6
- 238000005259 measurement Methods 0.000 claims description 6
- 230000007246 mechanism Effects 0.000 claims description 6
- 238000003058 natural language processing Methods 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 5
- 239000012530 fluid Substances 0.000 claims description 5
- 238000005452 bending Methods 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 claims description 3
- 238000009499 grossing Methods 0.000 claims description 3
- 239000000463 material Substances 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 230000007547 defect Effects 0.000 abstract description 7
- 238000012423 maintenance Methods 0.000 description 6
- 238000003825 pressing Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 239000002828 fuel tank Substances 0.000 description 3
- 239000007788 liquid Substances 0.000 description 3
- 238000007726 management method Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000013136 deep learning model Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000000737 periodic effect Effects 0.000 description 2
- 238000011176 pooling Methods 0.000 description 2
- 238000000926 separation method Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005672 electromagnetic field Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000011895 specific detection Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
-
- 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
- G01R31/003—Environmental or reliability tests
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
- G06F17/141—Discrete Fourier transforms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/20—Finite element generation, e.g. wire-frame surface description, tesselation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- Discrete Mathematics (AREA)
- Environmental & Geological Engineering (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Computer Graphics (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention relates to the technical field of fault detection, and discloses a transformer fault detection system based on an acoustic sensor array, which comprises a vibration noise characteristic analysis module, a noise characteristic signal enhancement module, an abnormal sound signal positioning module, a fault diagnosis module and a fault monitoring and early warning module; establishing a three-dimensional simulation model of transformer winding vibration by using coupling analysis software, realizing the enhancement of fault noise signals in an expected direction based on NAH reconstruction of two-dimensional Discrete Fourier Transform (DFT), performing performance analysis and evaluation based on a spatial stereo array of a far-field model multi-sound sensing array element, and positioning the abnormal sound position; and establishing a mode identification model of the transformer noise based on a signal processing algorithm. According to the invention, the radiation sound field of the transformer is measured through the sensor array, the acoustic image of the transformer is obtained, the mechanical condition of the equipment is evaluated, the defects of the traditional offline detection technology are avoided, and the non-intrusive live detection of the running condition of the transformer can be realized.
Description
Technical Field
The invention relates to the technical field of fault detection, in particular to a transformer fault detection system based on an acoustic sensor array.
Background
The power transformer has complex operation environment and high cost, plays a decisive role in stable operation of a power system, and once the power transformer is damaged due to faults, has extremely wide spread range and can cause casualties and major economic loss. Therefore, the evaluation of the running state of the transformer and the fault diagnosis technology are always hot spots of domestic and foreign research. According to statistics of related departments, more than 70% of transformers with accidents in China during 2000-2015 are subjected to mechanical faults, wherein the winding structure is subjected to unrecoverable changes such as winding loosening, warping, bulging, dislocation and the like under the action of electromagnetic force or mechanical force, and the conditions of iron core loosening, foreign body mechanical vibration, abnormal sound and the like are also included. Therefore, the detection of the mechanical state of the power transformer is very important to guarantee the safe and reliable operation of the power grid and improve the maintenance level of equipment.
At present, the common methods for monitoring the mechanical state of the transformer can be classified into an off-line detection method represented by a frequency response method and a short-circuit reactance method and a live detection method represented by a vibration acoustic signal analysis method. The off-line detection method is not influenced by the operating environment, test data between different phases can be compared with longitudinal historical data, the accuracy is high, and the method is mainly used for detecting the mechanical state of a winding component. The acoustic signal and the image forming analysis method thereof have attracted extensive attention by the advantages that the acoustic signal and the image forming analysis method thereof are not electrically connected with the tested transformer, the operation of the transformer is not influenced in the testing process, the acoustic signal contains more abundant mechanical states of the transformer, and the like, and the method for diagnosing the mechanical state of the winding by using the vibration signal is also applied on site, but still has the defect of difficult positioning.
Disclosure of Invention
The purpose of the invention is as follows: aiming at the defects of the prior art, the invention provides a transformer fault detection system based on an acoustic sensor array, which measures the radiation sound field of a transformer through the sensor array, obtains the acoustic image of the transformer, evaluates the mechanical condition of equipment, avoids the defects of the traditional offline detection technology, can realize the non-intrusive live detection of the running condition of the transformer, reduces the influence caused by the running quit of the transformer, and has obvious direct benefit.
The technical scheme is as follows: the invention provides a transformer fault detection system based on an acoustic sensor array, which comprises a vibration noise characteristic analysis module, a noise characteristic signal enhancement module, an abnormal sound signal positioning module, a fault diagnosis module and a fault monitoring and early warning module, wherein the vibration noise characteristic analysis module is used for analyzing the vibration noise characteristic of a transformer;
the vibration noise characteristic analysis module is used for establishing a three-dimensional simulation model of the vibration of the transformer winding, which comprises a winding, transformer oil, an oil tank and a fastener, based on coupling analysis software to obtain noise characteristics, mechanism and characteristic parameters;
the noise characteristic signal enhancement module is used for enhancing the noise characteristic of the transformer based on NAH reconstruction of two-dimensional Discrete Fourier Transform (DFT);
the abnormal sound signal positioning module is used for analyzing the acoustic environment of the periphery of the transformer substation and the transformer, analyzing and evaluating the performance of the spatial stereo array based on the multi-sound sensing array elements of the far field model and positioning the abnormal sound position;
the fault diagnosis module establishes a mode recognition model of transformer noise based on a signal processing algorithm, optimizes a noise extraction algorithm model by using noise voiceprint characteristics of the transformer under different working conditions, and confirms the selection, the number, the arrangement mode, the data acquisition and preprocessing method and the anti-interference measures of the transformer acoustic vibration sensors;
the fault monitoring and early warning module calculates and analyzes the noise characteristics of the transformer in real time and carries out voiceprint recognition and early warning through a transformer noise evaluation system of the high-performance data acquisition system, and constructs a semantic analysis technical framework of a transformer fault monitoring report through a Natural Language Processing (NLP) technology and by combining normative description of relevant specifications on transformer faults.
Furthermore, the coupling analysis software is a simulation technology which establishes a three-dimensional simulation model of the transformer winding vibration containing windings, transformer oil, an oil tank and fasteners, establishes electromagnetic-structure-fluid coupling analysis in the finite element software through three-dimensional finite element modeling, verifies and perfects the transformer winding vibration and the transmission characteristic thereof through a test means according to a simulation result, and obtains the transmission characteristic of normal and fault winding vibration, the vibration distribution rule of each point of the oil tank, the corresponding relation of the winding fault position and the transformer oil tank sound field radiation characteristic. Taking winding vibration as an example, the vibration of a winding of a power transformer in operation mainly comes from forced vibration caused by electromagnetic force of periodic action of the winding which is electrified with alternating current in an alternating magnetic field, the vibration of the winding is transmitted to the surface of an oil tank through transformer oil, an upper pressing plate, a lower pressing plate and a pressing nail of the winding so as to cause the vibration of the surface of the oil tank, the transmission of the transformer vibration in the oil meets a wave equation in three-dimensional liquid, the vibration of the winding has a large influence on the vibration of the adjacent tank wall from the wave equation and the transmission characteristic of vibration waves in the oil tank, and the vibration is gradually weakened along with the increase of the transmission distance, so that the measurement of the vibration in a certain area of the surface of the oil tank can effectively realize the correspondence between the vibration of the winding and the vibration of the tank wall, meanwhile, when the winding is deformed or the mechanical state is changed, the vibration of the winding can be changed to different degrees, and the change is transmitted to the surface of the oil tank through the oil to cause the change of the radiation sound field of the surface of the oil tank so as to react, therefore, the theory of vibration transmission and radiation of the oil tank is established, and the corresponding signal of a measuring point of the electromagnetic winding is guided to perfect by utilizing the electromagnetic-structure-fluid coupling of the finite element method to perfect the electromagnetic field.
Further, the noise characteristic signal enhancement module performs data observation by the following three methods:
firstly, zero padding is carried out around data, but the discontinuity of the data after zero padding can seriously influence the reconstruction result;
secondly, smoothing the data after zero padding, wherein in the traditional method, a proper window function is added to the data after zero padding, but the measured holographic surface information is distorted, so that a new reconstruction error is generated;
thirdly, extrapolating the measured data in a real space domain or a wave number domain, wherein the measurement parameters of the NAH are selected according to the size of a mechanical noise source, the frequency range and the requirements of imaging resolution, and the method mainly comprises the following steps: sampling frequency, size and layout of the microphone array, and distance of the microphone array from the sound source plane.
Furthermore, the abnormal sound signal positioning module is determined by testing a transformer acoustic vibration signal of the sensor array, specifically, according to a transmission mechanism of the transformer in a thin plate structure of the oil tank wall, vibration is mainly transmitted in the form of bending waves in the thin plate, which indicates that vibration on the surface of the oil tank has the characteristic of fluctuation, the wavelength and frequency of the vibration waves are related to the thickness of the oil tank wall and the young modulus of the material, the mechanical failure condition of the transformer is complex, different mechanical failures may occur simultaneously, and in addition, the internal vibration sources of the transformer are many, such as windings, iron cores and loose screws, by using the characteristics of vibration transmission and radiation sound field, when the winding state, the iron core state, the existence of foreign matters and the loose mechanical structure of the clamp are changed, the vibration signals will affect the oil tank wall and the radiated acoustic image, so that the array of the acoustic vibration sensors is constructed, the array formed by arranging the acoustic vibration sensor array according to a certain geometric structure, and the sensor array greatly explores useful information in the acoustic signals: such as signal direction recognition or source recognition, vibro-acoustic signal separation, vibro-acoustic signal de-noising and enhancement, vibro-acoustic camera applications. The image method is used for positioning the fault source, judging the position of the fault, and separating and judging the fault types of different fault source sound vibration signals, such as winding deformation looseness, iron core looseness and screw looseness, so that the functions of judging and positioning the mechanical fault of the transformer are finally realized.
Furthermore, the method for acquiring the oil tank radiation acoustic image by the acoustic vibration sensor array is used for clearing the change rule and the influence factors of the transformer oil tank surface radiation acoustic image under the normal mechanical condition and in various fault modes, acquiring the acoustic image characteristic parameters for diagnosing the mechanical state of the transformer, wherein the acoustic image is a colorful image formed by acoustic signals acquired by a plurality of acoustic vibration sensors, extracting the characteristic parameters in the image through image processing and recognition algorithms, and further forming related criteria.
Further, the fault monitoring and early warning module adopts a deep belief network to extract and classify characteristics of reconstructed sound pressure distribution, a convolutional neural network in a deep learning model is utilized to establish a multi-layer neural network model, as shown in fig. 3, acoustic image characteristics are extracted through typical steps of local perception, weight sharing and pooling by utilizing the convolutional neural network, as shown in fig. 4, an acoustic image is convoluted through three trainable filters and an applicable bias, three feature mapping maps are generated on a C1 layer, then four pixels of each group in the feature mapping maps are summed, weighted and biased, a Sigmoid function is used to obtain three feature mapping maps of an S2 layer, the mapping maps are further filtered to obtain a C3 layer, the hierarchical structure is used as S2 to generate S4, finally, the pixel values are rasterized and connected into a vector to be input to the traditional neural network, the obtained output convolutional network is essentially an input-to-output mapping, the convolutional network is trained by utilizing a known mode, the network has the mapping between transmission, the acoustic capability of the acoustic network during actual operation and the acoustic transformer during the operation are extracted through a fault diagnosis and parameter positioning method based on the convolutional network.
Advantageous effects
The transformer fault detection system based on the acoustic sensor array measures the radiation sound field of the transformer through the sensor array to obtain the acoustic image of the transformer, evaluates the mechanical condition of equipment, avoids the defects of the traditional off-line detection technology, can realize non-intrusive live detection of the operation condition of the transformer, reduces the influence caused by the fact that the transformer exits from operation, has obvious direct benefit, can effectively improve the operation reliability of the transformer by obtaining the acoustic image for diagnosis, provides reference for the non-intrusive live maintenance, on-line monitoring and maintenance departments of the equipment to timely master the mechanical state of the transformer, can make a targeted operation maintenance strategy according to the reference, avoids huge losses of national production and life caused by equipment faults, and can also avoid huge economic losses caused by power production and operation interruption.
Drawings
FIG. 1 is an overall system diagram of the present invention;
FIG. 2 is a diagram of a coupling analysis step according to the present invention;
FIG. 3 is a diagram of a model of a multi-layer neural network of the present invention;
FIG. 4 is a diagram of a convolutional neural network model of the present invention.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.
The invention discloses a transformer fault detection system based on an acoustic sensor array, which comprises a vibration noise characteristic analysis module, a noise characteristic signal enhancement module, an abnormal sound signal positioning module, a fault diagnosis module and a fault monitoring and early warning module.
The vibration noise characteristic analysis module is used for establishing a three-dimensional simulation model of the vibration of the transformer winding, which contains a winding, transformer oil, an oil tank and a fastener, by utilizing coupling analysis software.
The noise characteristic signal enhancement module is completed based on NAH reconstruction of two-dimensional discrete Fourier transform DFT, and the enhancement of the fault noise signal in the expected direction is realized.
The abnormal sound signal positioning module analyzes the acoustic environment of the periphery of the transformer substation and the transformer, and performs performance analysis and evaluation on the spatial stereo array with the multiple sound sensing array elements based on the far field model.
The fault diagnosis module establishes a mode recognition model of the transformer noise based on a signal processing algorithm, realizes intelligent extraction and reliable diagnosis of transformer fault characteristics, optimizes a noise extraction algorithm model by using noise voiceprint characteristics of the transformer under different working conditions, and improves accuracy and effectiveness of transformer noise fault diagnosis.
The fault monitoring and early warning module is used for calculating and analyzing the noise characteristics of the transformer in real time and carrying out voiceprint recognition and early warning by using a transformer noise evaluation system of a high-performance data acquisition system, processing an NLP (non line processing) technology by using a natural language, and constructing a semantic analysis technology framework of a transformer fault monitoring report by combining the normative description of relevant specifications on the transformer fault so as to realize the automatic recognition and digital management of operation and maintenance information.
For the above-mentioned transformer fault detection system based on acoustic sensor array, the specific detection management method is as follows:
101. firstly, a vibration noise characteristic analysis module utilizes coupling analysis software to carry out three-dimensional simulation modeling on a transformer so as to obtain noise characteristics, mechanisms and characteristic parameters.
In this embodiment, it is specifically explained that the coupling analysis software is a simulation technique that establishes a three-dimensional simulation model of transformer winding vibration including a winding, transformer oil, an oil tank, and a fastener, establishes electromagnetic-structure-fluid coupling analysis in the finite element software by using three-dimensional finite element modeling, and verifies and perfects the transformer winding vibration and its transmission characteristics by using a test means according to the simulation result, so as to obtain the transmission characteristics of normal and fault winding vibration, the vibration distribution rules of each point of the oil tank, the correspondence of winding fault positions, and the transformer oil tank sound field radiation characteristics. Taking winding vibration as an example, the vibration of a winding of a power transformer in operation mainly comes from forced vibration caused by electromagnetic force of periodic action of the winding which is electrified with alternating current in an alternating magnetic field, the vibration of the winding is transmitted to the surface of an oil tank through transformer oil, an upper pressing plate, a lower pressing plate, a pressing nail and the like of the winding so as to cause the vibration of the surface of the oil tank, at present, research on the vibration mechanism and the characteristic of the winding is carried out completely and abundantly, but for the transmission mode of the winding vibration, the energy proportion of the liquid and solid transmission vibration is not clear, so that the winding mechanical fault diagnosis and evaluation method based on the vibration method lacks an important theoretical basis, the principle of the method is not strong, only the transformer is used as a black box for diagnosis in use, and actually, the transmission of the transformer vibration in the oil meets the fluctuation equation in three-dimensional liquid, from the view of wave equation and the propagation characteristic of vibration wave in the oil tank, the vibration of the winding has larger influence on the vibration of the adjacent tank wall, and the vibration is gradually weakened along with the increase of the propagation distance, which shows that the measurement of the vibration in a certain area of the surface of the oil tank can effectively realize the correspondence between the vibration of the winding and the vibration of the tank wall, meanwhile, when the winding is deformed or the mechanical state is changed, the vibration of the winding can be changed in different degrees, and the change is transmitted to the surface of the oil tank through oil to further cause the change of the radiation sound field of the surface of the oil tank to react, therefore, the theory of vibration transmission and sound field radiation is established, and the electromagnetic-structure-fluid coupling of a finite element method is utilized to perfect the theory, guide the arrangement of the measuring points of a microphone, and establish the corresponding relation between the radiation signal of the surface of the oil tank and the fault position of the winding, the present embodiment is not particularly limited.
102. And then, enhancing the noise characteristics of the transformer by a noise characteristic signal enhancement module, and positioning the abnormal sound position by an intention signal positioning module.
In this embodiment, it is specifically explained that the noise characteristic signal enhancement module is completed by NAH reconstruction based on two-dimensional discrete fourier transform DFT, and the main reasons include: firstly, the transformer has a complex structure, abnormal noise emitted in a fault state is weak, evanescent wave components can be captured by adopting near-field measurement through NAH, and higher reconstruction accuracy can be obtained; the NAH is used for diagnosing the mechanical fault of the transformer for the first time, the NAH of the two-dimensional discrete Fourier transform DFT is the most applied NAH method at present, and the use method is skilled; thirdly, the defects of the NAH based on DFT can be compensated by the pretreatment of the measured data.
Specifically, the noise characteristic signal enhancement module performs data observation by the following three methods: firstly, zero padding is carried out around data, but the discontinuity of the data after zero padding can seriously influence the reconstruction result; secondly, smoothing the data after zero padding, wherein in the traditional method, a proper window function is added to the data after zero padding, but the measured holographic surface information is distorted, so that a new reconstruction error is generated; thirdly, extrapolating the measured data in a real space domain or a wave number domain, wherein the measurement parameters of the NAH are selected according to the size of a mechanical noise source, the frequency range and the requirements of imaging resolution, and the method mainly comprises the following steps: the sampling frequency, the size and layout of the microphone array, and the distance between the microphone array and the sound source plane are not particularly limited in this embodiment.
The abnormal sound signal positioning module is used for judging by utilizing a transformer sound vibration signal test of a sensor array, specifically, according to a transmission mechanism of a transformer in a thin plate structure of a fuel tank wall, vibration is mainly transmitted in a bending wave form in the thin plate, so that the vibration of the fuel tank surface has the characteristic of fluctuation, the wavelength and the frequency of the vibration wave are related to the thickness of the fuel tank wall and the Young modulus of a material, the mechanical fault condition of the transformer is complex, different mechanical faults can occur simultaneously, in addition, the vibration sources in the transformer are more, such as windings, iron cores, loose screws and the like, the transmission path is complex and other reasons cause that a single-point vibration signal cannot accurately judge the fault, and the existing vibration signal analysis method is a diagnosis and analysis method aiming at the surface of the transformer and a single point or a few points, the connection between the points is not considered, and the obtained vibration signal information is not fully utilized; in addition, the currently widely developed research only utilizes vibration signals, but the utilization of radiated acoustic image information is not sufficient, and if the method is assisted with a noise image, the method may be more helpful for the identification and positioning of sound source signals, and a method for measuring the acoustic vibration signals of the transformer based on the sensor array is proposed, according to the characteristics of vibration transmission and radiation sound field, when the mechanical structures such as winding state, iron core state, existence of foreign matters, clamp loosening and the like are changed, the acoustic vibration signals will be affected on the wall of the oil tank and the radiated acoustic images, therefore, the array of the acoustic vibration sensors is constructed, the form of the sensor array is like an acoustic array, the array which is formed by arranging according to a certain geometric structure has strong space selectivity, the source signals can be directionally enhanced, automatically monitored, positioned and tracked without mechanical movement on the array, the radiation sound field of the transformer is regarded as a plane wave, and the common acoustic sensor array is designed as a one-dimensional linear array or a two-dimensional array such as an equiregular rectangular grid array, a cross-axis array, a circular ring array, a non-regular archimedean array, a wheel array, and a fan-shaped acoustic sensor array, and a great useful signal can be discovered: such as signal direction recognition or source recognition, acoustic-vibration signal separation, acoustic-vibration signal de-noising and enhancement, and application of an acoustic-vibration camera, the image method is used for positioning a fault source, judging the position of the fault, and separating and judging fault types, such as winding deformation and looseness, iron core looseness, screw looseness and the like, of different fault sources, so as to finally realize the functions of judging and positioning the mechanical fault of the transformer, and the embodiment is not particularly limited.
103. And confirming the type selection, the number and the arrangement mode of the transformer sound vibration sensors, a data acquisition and preprocessing method and anti-interference measures by using a fault diagnosis module.
In this embodiment, it is specifically explained that the fault diagnosis module establishes a mode identification model of transformer noise based on a signal processing algorithm, so as to achieve intelligent extraction and reliable diagnosis of transformer fault characteristics, optimizes a noise extraction algorithm model by using noise voiceprint characteristics of a transformer under different working conditions, and improves accuracy and effectiveness of transformer noise fault diagnosis, for example, a method for obtaining an acoustic image of an oil tank by using an acoustic imaging technology performs tests on different operation working conditions of a winding, such as an oil tank radiation sound field under different load currents and temperatures, and compares the test results with simulation results, and records and stores transformer fault reasons for typical mechanical faults and defects of a transformer, such as winding faults, iron core faults, screw loosening abnormal sounds, and the like, by artificially setting faults in a transformer and a change rule of the oil tank radiation sound field, and the specific steps are shown in fig. 1.
104. And finally, calculating and analyzing the noise characteristics of the transformer in real time and performing voiceprint recognition and early warning by using a transformer noise evaluation system of the high-performance data acquisition system by using a fault monitoring and early warning module, processing an NLP (non-line-of-sight) technology by using a natural language, and constructing a semantic analysis technical framework of a transformer fault monitoring report by combining the normative description of relevant specifications on the transformer fault, thereby realizing the automatic recognition and digital management of operation and maintenance information.
In this embodiment, it is specifically explained that the fault monitoring and early warning module extracts and classifies features of the reconstructed sound pressure distribution by using a deep belief network, establishes a multi-layer neural network model by using a convolutional neural network in a deep learning model, as shown in fig. 3, extracts acoustic image characteristics by using the convolutional neural network through typical steps of local sensing, weight sharing, pooling, and the like, as shown in fig. 4, convolves an acoustic image with three trainable filters and an applicable bias, generates three feature maps at a C1 layer, then sums up four pixels of each group in the feature maps, weights, and biases, obtains three feature maps at an S2 layer by using a Sigmoid function, and filters the maps to obtain a C3 layer, and generates S4 in the same manner as S2, and finally, the pixel values are rasterized and connected into a vector to be input to the neural network, the obtained output convolution network is essentially a mapping from input to output, the convolution network is trained by a known mode, the network has the mapping capability between transmission, characteristic parameters are extracted from the sum acoustic images of a plurality of transformers in actual operation through the convolution neural network to form a transformer mechanical fault diagnosis and positioning method based on the acoustic imaging technology, the change rule and the influence factors of the transformer oil tank surface radiation acoustic images under the normal mechanical condition and various fault forms are clarified through a method of obtaining the oil tank radiation acoustic images through a sound vibration sensing array, the acoustic image characteristic parameters for diagnosing the mechanical state of the transformer are obtained, the acoustic image is a colorful image consisting of acoustic signals collected by a plurality of sound vibration sensors, the characteristic parameters in the image are extracted through an image processing and recognition algorithm, thereby forming a relevant criterion.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (6)
1. A transformer fault detection system based on an acoustic sensor array is characterized in that: the system comprises a vibration noise characteristic analysis module, a noise characteristic signal enhancement module, an abnormal sound signal positioning module, a fault diagnosis module and a fault monitoring and early warning module;
the vibration noise characteristic analysis module is used for establishing a three-dimensional simulation model of the vibration of the transformer winding, which comprises a winding, transformer oil, an oil tank and a fastener, based on coupling analysis software to obtain noise characteristics, mechanism and characteristic parameters;
the noise characteristic signal enhancement module is used for enhancing the noise characteristic of the transformer based on NAH reconstruction of two-dimensional Discrete Fourier Transform (DFT);
the abnormal sound signal positioning module is used for analyzing the acoustic environment of the periphery of the transformer substation and the transformer, analyzing and evaluating the performance of the spatial stereo array based on the multi-sound sensing array elements of the far field model and positioning the abnormal sound position;
the fault diagnosis module establishes a mode recognition model of transformer noise based on a signal processing algorithm, optimizes a noise extraction algorithm model by using noise voiceprint characteristics of the transformer under different working conditions, and confirms the selection, the number, the arrangement mode, the data acquisition and preprocessing method and the anti-interference measures of the transformer acoustic vibration sensors;
the fault monitoring and early warning module calculates and analyzes the noise characteristics of the transformer in real time and carries out voiceprint recognition and early warning through a transformer noise evaluation system of the high-performance data acquisition system, and constructs a semantic analysis technical framework of a transformer fault monitoring report through a Natural Language Processing (NLP) technology and by combining normative description of relevant specifications on transformer faults.
2. The system of claim 1, wherein the transformer fault detection system based on the acoustic sensor array comprises: the coupling analysis software is a simulation technology which is used for establishing a three-dimensional simulation model of transformer winding vibration containing a winding, transformer oil, an oil tank and a fastener, establishing electromagnetic-structure-fluid coupling analysis in the finite element software through three-dimensional finite element modeling, verifying and perfecting the transformer winding vibration and the transmission characteristic thereof through a test means according to a simulation result, and obtaining the transmission characteristic of normal and fault winding vibration, the vibration distribution rule of each point of the oil tank, the corresponding relation of winding fault positions and the transformer oil tank sound field radiation characteristic.
3. The system of claim 1, wherein the transformer fault detection system based on the acoustic sensor array comprises: the noise characteristic signal enhancement module carries out observation data by the following three methods:
first, zero padding is performed around the data;
secondly, smoothing the data after zero padding;
thirdly, extrapolating the measured data in a real space domain or a wave number domain, wherein the measurement parameters of the NAH are selected according to the size of a mechanical noise source, the frequency range and the imaging resolution, and the method comprises the following steps: sampling frequency, size and layout of the microphone array, and distance of the microphone array from the sound source plane.
4. The system of claim 1, wherein the system comprises: the abnormal sound signal positioning module judges by testing a sound vibration signal of the transformer of the sensor array, vibration is transmitted in a bending wave mode in a thin plate, vibration on the surface of an oil tank has the characteristic of fluctuation, the wavelength and the frequency of the vibration wave, the thickness of the wall of the oil tank and the Young modulus of a material are related, a vibration source in the transformer comprises a winding, an iron core and a loose screw, the sound vibration sensor array is constructed by utilizing the characteristics of vibration transmission and radiation sound field, and the array is formed by arranging the sound vibration sensor array according to a geometric structure.
5. The system of claim 4, wherein the transformer fault detection system based on the acoustic sensor array comprises: the method for acquiring the oil tank radiation acoustic image by the acoustic vibration sensing array is used for clearing the change rule and the influence factors of the transformer oil tank surface radiation acoustic image under the condition of normal mechanical condition and various fault forms, acquiring the acoustic image characteristic parameters for diagnosing the mechanical state of the transformer, and extracting the characteristic parameters in the image by image processing and recognition algorithms.
6. The system of claim 1, wherein the transformer fault detection system based on the acoustic sensor array comprises: the fault monitoring and early warning module adopts a deep belief network to extract and classify the characteristics of the reconstructed sound pressure distribution, acoustic images are convoluted with three trainable filters and can be biased, three characteristic mapping maps are generated on a C1 layer, four pixels of each group in the characteristic mapping maps are summed, weighted values are added with bias, three characteristic mapping maps of an S2 layer are obtained through a Sigmoid function, the mapping maps are filtered to obtain a C3 layer, a hierarchical structure and S2 layer generate S4, pixel values are rasterized and connected into a vector to be input into a neural network, the output convolutional network is essentially an input-to-output mapping, the convolutional network is trained by using a known mode, the characteristic parameters of the sum acoustic images of a plurality of transformers in actual operation are extracted through the convolutional neural network, and the transformer mechanical fault diagnosis and positioning method based on the acoustic imaging technology is formed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210876670.4A CN115358110A (en) | 2022-07-25 | 2022-07-25 | Transformer fault detection system based on acoustic sensor array |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210876670.4A CN115358110A (en) | 2022-07-25 | 2022-07-25 | Transformer fault detection system based on acoustic sensor array |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115358110A true CN115358110A (en) | 2022-11-18 |
Family
ID=84032760
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210876670.4A Pending CN115358110A (en) | 2022-07-25 | 2022-07-25 | Transformer fault detection system based on acoustic sensor array |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115358110A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117031154A (en) * | 2023-08-07 | 2023-11-10 | 国网山西省电力公司超高压变电分公司 | Transformer fault analysis method and system based on voiceprint recognition |
| CN117169639A (en) * | 2023-11-02 | 2023-12-05 | 启东市航新实用技术研究所 | Product detection method and system for power adapter production |
| CN117633504A (en) * | 2023-10-30 | 2024-03-01 | 国网江苏省电力有限公司淮安供电分公司 | Optical fiber sensing evaluation method and device for state of oil immersed transformer |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101865789A (en) * | 2010-06-30 | 2010-10-20 | 上海交通大学 | Fault detecting device of near field acoustic holography sound image mode identification and detecting method thereof |
| CN107748314A (en) * | 2017-10-18 | 2018-03-02 | 国网重庆市电力公司北碚供电分公司 | Transformer Faults Analysis system based on sound wave shock detection |
| CN108760034A (en) * | 2018-05-21 | 2018-11-06 | 广西电网有限责任公司电力科学研究院 | A kind of transformer vibration noise source positioning system and method |
| CN109033612A (en) * | 2018-07-20 | 2018-12-18 | 广西电网有限责任公司电力科学研究院 | A kind of Diagnosis Method of Transformer Faults based on vibration noise and BP neural network |
| CN110514924A (en) * | 2019-08-12 | 2019-11-29 | 武汉大学 | The Winding in Power Transformer Fault Locating Method of depth convolutional neural networks fusion visual recognition |
| CN110534118A (en) * | 2019-07-29 | 2019-12-03 | 安徽继远软件有限公司 | Transformer/reactor method for diagnosing faults based on Application on Voiceprint Recognition and neural network |
| CN111175698A (en) * | 2020-01-18 | 2020-05-19 | 国网山东省电力公司菏泽供电公司 | Transformer noise source positioning method, system and device based on sound and vibration combination |
| CN111537822A (en) * | 2020-05-22 | 2020-08-14 | 国网上海市电力公司 | Transformer fault diagnosis system based on acoustic imaging |
| CN113707176A (en) * | 2021-09-02 | 2021-11-26 | 国网安徽省电力有限公司铜陵供电公司 | Transformer fault detection method based on acoustic signal and deep learning technology |
| CN114235366A (en) * | 2021-12-14 | 2022-03-25 | 国网天津市电力公司电力科学研究院 | Transformer fault identification method based on transformer noise vibration characteristics |
| CN114547924A (en) * | 2021-11-12 | 2022-05-27 | 国网江苏省电力有限公司南京供电分公司 | Multi-physical-field-coupling-based transformer winding vibration voiceprint analysis method and system |
-
2022
- 2022-07-25 CN CN202210876670.4A patent/CN115358110A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101865789A (en) * | 2010-06-30 | 2010-10-20 | 上海交通大学 | Fault detecting device of near field acoustic holography sound image mode identification and detecting method thereof |
| CN107748314A (en) * | 2017-10-18 | 2018-03-02 | 国网重庆市电力公司北碚供电分公司 | Transformer Faults Analysis system based on sound wave shock detection |
| CN108760034A (en) * | 2018-05-21 | 2018-11-06 | 广西电网有限责任公司电力科学研究院 | A kind of transformer vibration noise source positioning system and method |
| CN109033612A (en) * | 2018-07-20 | 2018-12-18 | 广西电网有限责任公司电力科学研究院 | A kind of Diagnosis Method of Transformer Faults based on vibration noise and BP neural network |
| CN110534118A (en) * | 2019-07-29 | 2019-12-03 | 安徽继远软件有限公司 | Transformer/reactor method for diagnosing faults based on Application on Voiceprint Recognition and neural network |
| CN110514924A (en) * | 2019-08-12 | 2019-11-29 | 武汉大学 | The Winding in Power Transformer Fault Locating Method of depth convolutional neural networks fusion visual recognition |
| US20210048487A1 (en) * | 2019-08-12 | 2021-02-18 | Wuhan University | Power transformer winding fault positioning method based on deep convolutional neural network integrated with visual identification |
| CN111175698A (en) * | 2020-01-18 | 2020-05-19 | 国网山东省电力公司菏泽供电公司 | Transformer noise source positioning method, system and device based on sound and vibration combination |
| CN111537822A (en) * | 2020-05-22 | 2020-08-14 | 国网上海市电力公司 | Transformer fault diagnosis system based on acoustic imaging |
| CN113707176A (en) * | 2021-09-02 | 2021-11-26 | 国网安徽省电力有限公司铜陵供电公司 | Transformer fault detection method based on acoustic signal and deep learning technology |
| CN114547924A (en) * | 2021-11-12 | 2022-05-27 | 国网江苏省电力有限公司南京供电分公司 | Multi-physical-field-coupling-based transformer winding vibration voiceprint analysis method and system |
| CN114235366A (en) * | 2021-12-14 | 2022-03-25 | 国网天津市电力公司电力科学研究院 | Transformer fault identification method based on transformer noise vibration characteristics |
Non-Patent Citations (3)
| Title |
|---|
| 林春清 等: "面向电力变压器的声纹智能诊断装置设计与应用", 传感技术学报, vol. 34, no. 10, pages 1412 - 1420 * |
| 王旭平 等: "机电设备检测与诊断技术", 西北工业大学出版社, pages: 130 - 139 * |
| 邵宇鹰 等: "基于近场声全息的电力变压器噪声源识别***", 华东电力, vol. 42, no. 12, pages 2522 - 2524 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117031154A (en) * | 2023-08-07 | 2023-11-10 | 国网山西省电力公司超高压变电分公司 | Transformer fault analysis method and system based on voiceprint recognition |
| CN117031154B (en) * | 2023-08-07 | 2024-03-22 | 国网山西省电力公司超高压变电分公司 | Transformer fault analysis method and system based on voiceprint recognition |
| CN117633504A (en) * | 2023-10-30 | 2024-03-01 | 国网江苏省电力有限公司淮安供电分公司 | Optical fiber sensing evaluation method and device for state of oil immersed transformer |
| CN117633504B (en) * | 2023-10-30 | 2024-06-21 | 国网江苏省电力有限公司淮安供电分公司 | Optical fiber sensing evaluation method and device for state of oil immersed transformer |
| CN117169639A (en) * | 2023-11-02 | 2023-12-05 | 启东市航新实用技术研究所 | Product detection method and system for power adapter production |
| CN117169639B (en) * | 2023-11-02 | 2023-12-29 | 启东市航新实用技术研究所 | Product detection method and system for power adapter production |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115358110A (en) | Transformer fault detection system based on acoustic sensor array | |
| CN109033612B (en) | Transformer fault diagnosis method based on vibration noise and BP neural network | |
| WO2021232655A1 (en) | Method for evaluating mechanical state of high-voltage shunt reactor on the basis of vibration feature | |
| CN107748314A (en) | Transformer Faults Analysis system based on sound wave shock detection | |
| CN107036817B (en) | SVR rolling bearing performance decline prediction technique based on krill group's algorithm | |
| CN110751073A (en) | Pipeline early damage mode identification method based on distributed optical fiber sensing and deep learning | |
| CN109827777A (en) | Rolling bearing fault prediction technique based on Partial Least Squares extreme learning machine | |
| CN207300606U (en) | Transformer information collecting device based on Principles of Acoustics | |
| CN110728257B (en) | Transformer winding fault monitoring method based on vibration gray level image | |
| KR20200014129A (en) | Diagnosis method of electric transformer using Deep Learning | |
| CN111879397A (en) | Fault diagnosis method for energy storage mechanism of high-voltage circuit breaker | |
| CN109649432A (en) | Cloud platform rail integrity monitoring systems and method based on guided wave technology | |
| CN114217149A (en) | Transformer acoustic fingerprint uninterrupted power detection and state early warning method | |
| CN110632484A (en) | ELM-based GIS partial discharge defect diagnosis and classification system and method | |
| Chen et al. | Fault diagnosis of full-hydraulic drilling rig based on RS–SVM data fusion method | |
| CN117607598A (en) | Transformer fault detection method and system based on voiceprint characteristics | |
| CN113109666A (en) | Track circuit fault diagnosis method based on deep convolutional neural network | |
| CN116884432A (en) | VMD-JS divergence-based power transformer fault voiceprint diagnosis method | |
| CN117116293A (en) | Machine equipment fault diagnosis system in complex sound field environment | |
| CN116699400A (en) | Generator rotor short-circuit fault monitoring system, method and readable storage medium | |
| Zhao et al. | Damage detection of wind turbine blade based on wavelet analysis | |
| WO2023004180A1 (en) | Identification of false transformer humming using machine learning | |
| CN115406522A (en) | Power plant equipment running state research and application based on voiceprint recognition | |
| CN115468751A (en) | Method and device for sound collection and defect identification of transformer | |
| CN111983398A (en) | Power transformer partial discharge TDOA (time difference of arrival) positioning method and system under condition of considering NLOS (non line of sight) |
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 |