CN112395959B - Power transformer fault prediction and diagnosis method and system based on audio frequency characteristics - Google Patents
Power transformer fault prediction and diagnosis method and system based on audio frequency characteristics Download PDFInfo
- Publication number
- CN112395959B CN112395959B CN202011187629.3A CN202011187629A CN112395959B CN 112395959 B CN112395959 B CN 112395959B CN 202011187629 A CN202011187629 A CN 202011187629A CN 112395959 B CN112395959 B CN 112395959B
- Authority
- CN
- China
- Prior art keywords
- power transformer
- fault
- audio
- mel
- frequency
- 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
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000003745 diagnosis Methods 0.000 title claims abstract description 38
- 230000005236 sound signal Effects 0.000 claims abstract description 47
- 239000002245 particle Substances 0.000 claims abstract description 44
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 30
- 230000000739 chaotic effect Effects 0.000 claims abstract description 27
- 238000001228 spectrum Methods 0.000 claims abstract description 16
- 238000012549 training Methods 0.000 claims abstract description 13
- 238000012847 principal component analysis method Methods 0.000 claims abstract description 7
- 230000000737 periodic effect Effects 0.000 claims description 20
- 238000004364 calculation method Methods 0.000 claims description 17
- 238000005457 optimization Methods 0.000 claims description 15
- 238000013507 mapping Methods 0.000 claims description 14
- 230000006870 function Effects 0.000 claims description 13
- 239000011159 matrix material Substances 0.000 claims description 9
- 238000005516 engineering process Methods 0.000 claims description 8
- 238000012706 support-vector machine Methods 0.000 claims description 8
- 230000008859 change Effects 0.000 claims description 7
- 238000009826 distribution Methods 0.000 claims description 6
- 239000012634 fragment Substances 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 238000001514 detection method Methods 0.000 claims description 5
- 238000001914 filtration Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 4
- 108010074864 Factor XI Proteins 0.000 claims description 3
- 230000005540 biological transmission Effects 0.000 claims description 3
- 230000001186 cumulative effect Effects 0.000 claims description 3
- 238000013016 damping Methods 0.000 claims description 3
- 230000005284 excitation Effects 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 230000035772 mutation Effects 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 238000012546 transfer Methods 0.000 claims description 3
- 230000001131 transforming effect Effects 0.000 claims description 3
- 230000007704 transition Effects 0.000 claims description 3
- 238000009827 uniform distribution Methods 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 2
- 230000003862 health status Effects 0.000 claims 1
- 238000012423 maintenance Methods 0.000 description 7
- 230000009466 transformation Effects 0.000 description 7
- 238000011160 research Methods 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 3
- 230000036541 health Effects 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 230000008439 repair process Effects 0.000 description 3
- 230000035945 sensitivity Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000005312 nonlinear dynamic Methods 0.000 description 2
- 238000010079 rubber tapping Methods 0.000 description 2
- 230000003595 spectral effect Effects 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 238000004804 winding Methods 0.000 description 2
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical group [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000007792 addition Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 210000005069 ears Anatomy 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- WABPQHHGFIMREM-UHFFFAOYSA-N lead(0) Chemical compound [Pb] WABPQHHGFIMREM-UHFFFAOYSA-N 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000010183 spectrum analysis Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000000725 suspension Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H17/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves, not provided for in the preceding groups
-
- 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
- G01R31/50—Testing of electric apparatus, lines, cables or components for short-circuits, continuity, leakage current or incorrect line connections
- G01R31/62—Testing of transformers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Medical Informatics (AREA)
- Power Engineering (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Complex Calculations (AREA)
Abstract
The utility model provides a power transformer fault prediction and diagnosis method based on audio characteristics, which specifically comprises the following steps: s1, detecting effective signals of audio data of a power transformer under a noise background based on a chaotic oscillator; s2, calculating a logarithmic energy spectrum on a nonlinear Mel scale, and taking the logarithmic energy spectrum as a characteristic quantity of an audio signal of the power transformer; s3, calculating principal components of the characteristic quantity of the audio signal of the power transformer by adopting a principal component analysis method; s4, optimizing an optimal super-parameter training power transformer fault prediction model of a vector machine algorithm by adopting a quantum particle swarm algorithm; s5, if the power transformer is in a fault state, extracting fault characteristic frequency range amplitude values by adopting a 1/3 octave algorithm, comparing the fault characteristic frequency range amplitude values with an expert experience rule base, and predicting/obtaining the fault type of the power transformer. The utility model can improve the recognition precision of the power transformer running state fault prediction and reduce the calculated amount; fault determination may also be based on other metrology data or infrared images.
Description
Technical Field
The utility model belongs to the technical field of power transformer fault diagnosis, and particularly relates to a power transformer fault prediction and diagnosis method and system based on audio characteristics.
Background
In the power system, the power transformer plays important roles in voltage conversion, electric energy distribution, voltage regulation, isolation and the like, and is related to safe, stable, reliable and economic operation of the power system. The power transformer is complex in structure and mainly comprises an iron core, a winding, an oil tank, a conservator, an insulating sleeve, a tapping joint and the like, and is generally installed outdoors, the working environment is bad, faults are inevitably generated along with the increase of the running time, and the faults relate to the winding, the main insulation, a lead wire, a tapping switch, the sleeve and the like. The power department carries out primary overhaul within 5 years according to DL/T573-95 'power transformer overhaul guide rules' and the operation environment of the power transformer and 1) the power transformer newly put into operation, and then carries out primary overhaul every 10 years later; 2) Performing small repair at least once every year; "the standard schedules annual maintenance schedules for power transformers and the manner in which power systems operate. The power transformer is regularly overhauled, so that the effective utilization rate of the power transformer is easy to be reduced, a large amount of manpower and material resources are wasted, and even maintenance faults can be caused, and 'over repair or under repair' is generated.
In the running process of the power transformer, mechanical deformation is generated under the action of internal current and magnetic field, the vibration signal is expressed by self conduction, and the vibration signal is transmitted through an air medium, so that the sound signal of the power transformer in the running state is changed. The human ear can hear 20-20 kHz sound, and the experienced operation and maintenance personnel can judge whether the equipment is in a fault state or not by means of the ear hearing the sound of the running power transformer. The fault diagnosis method based on experience and subjective judgment of operation and maintenance personnel has strong uncertainty, the sound signals of the power transformer in the running state are obtained in real time based on the sound sensor device, various faults of the power transformer can be found in time by extracting the audio characteristics of the power transformer in the running state, prevention is achieved, convenience is provided for running management of the power transformer, basis is provided for maintenance, and waste of manpower and material resources is reduced. Therefore, developing a power transformer fault prediction and diagnosis system based on audio features has great social significance and economic benefit.
Studies on power transformer faults generally involve two directions: and (5) predicting and diagnosing faults. The fault diagnosis is to accurately diagnose the fault cause of the equipment after the equipment fails and give a specific processing method, but is only applicable to the equipment after the equipment fails. For important equipment which runs for a long time with high strength, the fault prediction can reduce maintenance cost, and the running condition of the equipment can be predicted in advance to reduce economic loss, so that maintenance is carried out in a planned way, and normal production is ensured. Therefore, the power transformer is indispensable to fault diagnosis and prediction, so that damage can be timely stopped and the power transformer is prevented from suffering.
The running sound of the power transformer can reflect the health state of the current equipment. Document 1 (Du Yiming. Research on a transformer fault diagnosis system based on acoustic signals [ D ]. University of science and technology, 2013.) has studied the generation mechanism of sound in the operating state of a power transformer, and has pointed out that whether the power transformer is in a normal operating state or a fault state, the sound frequency of the power transformer is basically within 1000Hz, and sensitivity of the sensor in a low frequency region and flatness of a frequency response curve should be considered when selecting the sensor; document 2 (Wu Song. Transformer fault diagnosis research based on acoustic characteristics [ D ]. University of science and technology in china, 2012.) demonstrates that the characteristic frequency of a transformer discharge sound signal is 250Hz, and the effect of an algorithm such as a spectrum analysis method, a wavelet algorithm, a Hi lbert-Huang transform and the like on signal extraction is comparatively analyzed; document 3 (Zhang Ruiqi. Research on a transformer discharge fault diagnosis method based on acoustic signals [ D ]. University of science and technology in China, 2018.) uses wavelet packets to analyze the energy distribution of sound signals in a normal operation state of a power transformer and a spark discharge state, and the analysis shows that 90% of sound energy in the normal operation state is distributed in a frequency band of 0-1280 Hz, the audio frequency characteristic frequency band of needle spark discharge is 5000-6000 Hz, the audio frequency characteristic frequency band of needle plate spark discharge is 6000-8000 Hz, and the audio frequency characteristic frequency band of suspension discharge is 1000-3000 Hz. Document 4 (Hou Zeng. Research on classification and positioning method of faults of power transformation equipment based on sound characteristics [ D ]. University of north-China electric power, 2018.) adopts a two-dimensional principal component analysis method to extract main frequency spectrum components in the running state of a power transformer, and positions a fault interval through a MUSIC algorithm; document 5 (Deng Kai. Extraction of characteristic parameters of discharge sound signals of transformer oil and recognition research [ D ]. University of science and technology, 2016.). The discharge sound signals of transformer oil are denoised by a wavelet threshold denoising method, and characteristic quantities of discharge fault sounds are extracted by a multi-scale characteristic entropy method; document 6 (Jin Xiao. Research on fault diagnosis method of distribution transformer based on acoustic signal [ D ]. University of martial arts, 2017.) adopts a rapid independent analysis algorithm based on negative entropy to separate a target sound source, and performs a complete set empirical mode decomposition algorithm on the target sound source, and extracts singular spectral entropy reflecting complexity and irregularity degree of the signal, complete set empirical mode frequency band energy entropy reflecting energy characteristics of the signal, and marginal spectral entropy reflecting time-frequency characteristics of the signal, and center of gravity frequency as characteristic quantities.
As can be seen from the summary analysis, the current power transformer fault prediction and diagnosis method based on the sound signal is researched by the following defects:
(1) The frequency domain analysis method is mostly adopted for extracting the operation sound characteristic quantity of the power transformer, or wavelet transformation, hilbert-Huang transformation and the like are adopted for extracting certain components of reaction faults, and further analysis is not carried out from the perspective of cepstrum and Mel cepstrum by referring to the recognition principle of human ears on sound;
(2) The power transformer has a complex structure, a severe running environment and non-negligible background noise, particularly when the background noise energy is large, the sound component reflecting the fault characteristics of the power transformer can be submerged, and most of the existing transformer fault prediction and diagnosis methods based on sound signals only adopt a simple filtering algorithm to filter noise;
(3) Because of limitations of the existing sound signal feature quantity extraction method, artificial intelligent methods such as a neural network and a Support Vector Machine (SVM) are not popular in the aspect of power transformer fault prediction based on sound signals.
Disclosure of Invention
Aiming at the problems, the utility model provides a power transformer fault prediction and diagnosis method and system based on audio characteristics, which can solve the problems in the prior art, monitor, predict and diagnose equipment faults in real time on line according to the audio characteristics, and have higher prediction precision when the fault characteristics are not obvious when submerged in noise; waste of manpower and material resources is reduced, and power supply reliability is improved; other metrology data for the power transformer may also be supported.
In order to achieve the above purpose, the present utility model adopts the following technical scheme:
a power transformer fault prediction and diagnosis method based on audio features comprises the following specific steps:
s1, detecting an effective signal of audio data of a power transformer under a strong noise background based on a chaotic oscillator;
s2, aiming at the audio effective signal extracted in the step S1, calculating a logarithmic energy spectrum on a nonlinear Mel scale by adopting Mel frequency cepstrum technology, and taking the logarithmic energy spectrum as a characteristic quantity of the audio signal of the power transformer; the Mel scale transformation is carried out in the extraction process, and the nonlinear transformation ensures that the sound signal has higher anti-noise performance;
s3, calculating principal component components of the audio signal characteristic quantity of the power transformer by adopting a PCA principal component analysis method;
s4, optimizing an optimal super-parameter training power transformer fault prediction model of an SVM vector machine algorithm by adopting a QPSO quantum particle swarm algorithm;
s5, if the power transformer is in a fault state, extracting fault characteristic frequency range amplitude values by adopting a 1/3 octave algorithm, calculating T2 and SEP statistics of an audio signal of the power transformer based on the 1/3 octave amplitude values, finding out a frequency range corresponding to an octave with the largest contribution to the T2 or SEP statistics when the T2 or SEP statistics exceeds a threshold value, serving as the fault audio characteristics of the power transformer, comparing the fault audio characteristics with an expert experience rule base, and predicting/obtaining the fault type of the power transformer.
Further, step S1 uses the power transformer audio signal as a weak periodic or quasi-periodic disturbance signal of the chaotic system, and uses the characteristics of the chaotic system that the phase state of the chaotic system is essentially changed due to perturbation and sensitivity of the chaotic system to parameters to detect the power transformer audio signal through calculation and identification. According to the nonlinear dynamics system, when the chaotic state and the large-scale periodic state are in the same state, the corresponding system phase state and the system state are different, and the transition from the chaotic phase state to the large-scale periodic phase state of the system is used as the detection basis of a weak periodic or quasi-periodic signal. The chaotic system is based on Duffing-Holmes, and the specific method comprises the following steps:
s101, the Duffing-Holmes oscillator equation is
Wherein k is the damping ratio, -x+x 3 For nonlinear restoring force term, rcos (ωt) is periodic strategy force or excitation signal, r is periodic strategy force amplitude, poincare mapping
Chaos in the sense of Smale horseshoes exists;
s102, calculating a threshold value for generating chaos by a Melnikov method, wherein R/k satisfies inequality R/k > R m
Wherein R is m For m harmonic rail bifurcation values of parameter R/k, when R/k crosses bifurcation value R m When the chaotic system changes in the topological property of the vector field in the phase space; when R/k is smaller than R m When the system is in a chaotic state, a periodic closed rail does not exist; when R/k is greater than R m When the system is in a large period state, a cluster of period closed tracks exist.
Further, the specific calculation method of the mel frequency cepstrum technique in the step S2 includes the following steps:
s201, pre-emphasis is carried out on the audio fragment by adopting a high-pass filter, so that the problem that the loss of a high-frequency component signal is larger than that of a low-frequency signal in the transmission process is solved:
H(z)＝1-bz -1
wherein b is 0.98;
s202, in order to avoid mutation at two end points during discrete FFT conversion, a Hamming window is adopted to carry out windowing treatment on the audio fragment:
w[n]＝0.54-0.46cos(2nπ/L)
wherein n represents a sampling point number, and L represents a window length;
s203, performing N-point discrete Fourier transform DFT on each frame of time domain signal, wherein N is greater than or equal to L:
s204, arranging a plurality of band-pass filters H in the sound frequency spectrum range m (k) M is more than or equal to 0 and less than M, wherein M is the number of filters;
the filter has triangular filtering characteristics, the center frequency is f (m), and the filter has equal bandwidth in the Mel frequency range; transforming the actual frequency domain into the Mel frequency domain by adopting the following method
m(f)＝1125ln(1+f/700)
Mapping the limited practical frequency range [ fl, fh ] to the Mel frequency range [ Fsell, fsell ] and equally dividing the Mel frequency range into M+1 parts to obtain M Mel center frequencies, and mapping the obtained M Mel center frequencies to the practical frequency domain to obtain the center frequencies of M triangular filters in the practical frequency domain;
according to the center frequency of the actual frequency domain, determining the transfer function of the Mel filter bank, wherein m is the filter sequence number:
the logarithmic energy of the frequency domain signal X (k) output by the mth Mel filter is calculated, and the expression is:
wherein E (m) is logarithmic energy, H m (k) For a mel filter bank, X (k) is the frequency domain signal;
s205, obtaining Mel cepstrum coefficient through discrete cosine transform DCT, wherein the expression is
Wherein: c (n) is the nth frequency cepstrum coefficient, E (M) is logarithmic energy, and M is the number of Mel filters, namely the output dimension.
Further, step S3 extracts principal components of the Mel cepstrum feature of the audio signal by using principal component analysis method, uses X to represent the Mel cepstrum feature matrix of the audio signal, R to represent the correlation coefficient matrix of X, and calculates the feature value according to the feature equation |r- λe|=0, i.e. solves the feature equation
r n λ m +r n-1 λ m-1 +…+r 1 λ+r 0 ＝0
Obtaining a characteristic value lambda 1 ,λ 2 ,…,λ m And arranging the characteristic values in order from large to small, i.e. lambda 1 ≥λ 2 ≥…≥λ m The method comprises the steps that (1) the principal component is determined according to a set cumulative contribution rate threshold value; wherein the contribution rate of the single feature isThe first k eigenvalues add up to contribute +.>Taking the first, second, … and p (p is less than or equal to m) main components corresponding to the characteristic values with the accumulated contribution reaching 85-95%.
Further, in step S4, the method for training the prediction model by using the vector machine algorithm includes:
given a training sample set of { (x) i ,y i ) I=1, 2, …, l }, where x i ∈R n Representing the input vector, y i E R represents the output result, and nonlinear mapping is adoptedInput vector x i Mapping to higher dimensional feature space R k (k>n) constructing an optimal hyperplane in the space>The "total deviation" of all sample points from the hyperplane is minimized,
wherein ω represents a weight coefficient vector; b is a bias constant;
using the epsilon-insensitive loss function, the deviation of the sample point from the optimal hyperplane can be expressed as
c(x,y,f(x))＝max(0,|y-f(x)|-ε)
Wherein ε represents the allowable error;
adding relaxation factor xi i 、When there is an error in the division, ζ i 、/>Are all greater than 0; when the division is error-free, xi i 、/>All take 0, convert to solve the optimization objective function minimization problem:
wherein, C represents penalty factor, and is converted into convex quadratic optimization problem; constraint optimization problem is solved by Lagrange coefficient method according to optimization condition KKT
Wherein alpha is i 、Representing Lagrange coefficients, SV representing the support vector.
Further, in step S4, the optimal super-parameters of the SVM are optimized by adopting a quantum particle swarm algorithm, the particle population size is set to be m, the particle dimension is set to be D, the position of each particle represents the current solution of the particle, and the ith particle has the following properties:
current position: x is x i ＝(x i1 ,x i2 ,…,x iD )；
Historical optimal position: p is p i ＝(p i1 ,p i2 ,…p iD )；
The average value of the historic optimal positions of all particles is
The next position of the particle is updated as follows
Wherein g best Is a global optimal position; alpha is a compression-expansion factor;and u is the value of the uniform distribution on (0, 1); the probability of taking the sum is 0.5.
Further, in step S5, 1/3 octave power spectrum-based method calculates electricityThe power transformer audio signal is windowed and framed according to the method described in step S202, and the 1/3 octave amplitude calculation result of each frame is recorded as x= (x) 1 ,x 2 ,…,x 29 ) And calculates T2 and SEP statistics of the power transformer audio signal based on the 1/3 octave amplitude,
T 2 statistics measure the change in principal component space of sample x:
wherein Λ= (λ) 1 ,λ 2 ,...,λ p )，Is a control limit with a confidence level of alpha;
the usual calculation method of the control limit is:
wherein F is α (p, n-p) is the F distribution value with p and n-p degrees of freedom with confidence α;
the SEP index measures the change in projection of the sample vector x in the residual space:
wherein,a control limit indicating a confidence level α;
the calculation formula of (2) is as follows:
wherein,i＝1,2,3，/> is the eigenvalue of the covariance matrix of X, C α A threshold value with the confidence degree alpha is distributed for standard normal;
based on T 2 The definition of the contribution graph of (a) is as follows:
the SPE-based contribution graph is defined as follows:
wherein,
when the T is detected 2 Or after the SEP statistic exceeds the threshold value, the frequency range corresponding to the octave with the largest contribution graph is regarded as the fault characteristic frequency, and the fault type possibly occurring in the power transformer is obtained through comparison with expert experience rules.
A power transformer fault prediction and diagnosis system based on audio features comprises a field device layer, a sensing device layer, an edge gateway layer, an AI big data platform and an external service layer; the sensing equipment layer uploads the real-time collected power transformer audio data to the edge gateway layer; the edge gateway layer predicts the running state of the power transformer according to the published prediction model and the received real-time audio data, and if the power transformer is in a fault state, the edge gateway uploads the audio data of the power transformer to the AI big data platform for fault diagnosis; the AI big data platform generates an alarm event after receiving the fault information, and simultaneously performs fault diagnosis on the power transformer according to the received real-time data; the external service layer provides a display page of the power transformer fault prediction and diagnosis system.
Furthermore, the AI big data platform can perform timing training and issue a prediction model, evaluate the health state of the power transformer at regular time, manage equipment, process data, serve data and control equipment urgently.
Further, the equipment sensing layer also uploads the power transformer temperature data acquired in real time and/or the infrared image to the edge gateway layer; fault detection and determination may also be made based on other measured data or infrared images of the power transformer.
Compared with the prior art, the utility model has the advantages that:
1. according to the characteristic that the perturbation and the sensibility of the chaotic system to parameters are the essential change of the phase state of the chaotic system, the effective audio signal under the running state of the power transformer is detected by adopting a calculation and identification method and is used for subsequent fault prediction;
2. according to the utility model, a Mel frequency cepstrum technology is adopted, the logarithmic energy spectrum of the extracted effective audio signal on the nonlinear Mel scale is calculated as the characteristic quantity of the audio signal, so that the recognition precision of fault prediction is improved;
3. according to the utility model, principal component of the cepstrum coefficient of the audio signal Mel is analyzed by adopting the PCA method, so that non-principal components are reduced, the prediction accuracy of the model is further improved, and the calculated amount is reduced;
4. according to the utility model, the super-parameters of the SVM are optimized by adopting a quantum particle swarm algorithm, so that the searching efficiency is effectively improved; when the power transformer is detected to be faulty, the type of the fault possibly occurring in the power transformer is estimated according to expert experience rules by calculating 1/3 octave of the audio signal of the power transformer, and a processing suggestion is given;
5. the power transformer fault prediction and diagnosis system not only supports the fault prediction and diagnosis of the running state of the power transformer based on the sound signal, but also supports the judgment based on other measurement data or infrared images of the power transformer.
Additional advantages, objects, and features of the utility model will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the utility model.
Drawings
FIG. 1 is a block diagram of a power transformer fault prediction and diagnostic system of the present utility model;
FIG. 2 is a flow chart of a power transformer fault prediction and diagnosis method of the present utility model;
FIG. 3 is a diagram of the Mel cepstrum calculation step of the present utility model;
FIG. 4 is a flow chart of a quantum particle swarm algorithm of the present utility model;
FIG. 5 is a flow chart of a 1/3 octave algorithm of the present utility model.
The device comprises a field device layer 1, a sensing device layer 2, an edge gateway layer 3, an AI big data platform 4 and an external service layer 5.
Detailed Description
The utility model will be described in further detail with reference to the drawings and the detailed description.
Example 1:
as shown in fig. 2, a power transformer fault prediction and diagnosis method based on audio features specifically includes the following steps:
s1, detecting an effective signal of audio data of a power transformer under a strong noise background based on a chaotic oscillator;
s2, aiming at the audio effective signal extracted in the step S1, calculating a logarithmic energy spectrum on a nonlinear Mel scale by adopting Mel frequency cepstrum technology, and taking the logarithmic energy spectrum as a characteristic quantity of the audio signal of the power transformer; the Mel scale transformation is carried out in the extraction process, and the nonlinear transformation ensures that the sound signal has higher anti-noise performance;
s3, calculating principal component components of the audio signal characteristic quantity of the power transformer by adopting a PCA principal component analysis method;
s4, optimizing an optimal super-parameter training power transformer fault prediction model of an SVM vector machine algorithm by adopting a QPSO quantum particle swarm algorithm;
s5, if the power transformer is in a fault state, extracting fault characteristic frequency range amplitude values by adopting a 1/3 octave algorithm, calculating T2 and SEP statistics of an audio signal of the power transformer based on the 1/3 octave amplitude values, and finding out T 2 Or the SEP statistic exceeds a threshold value for T 2 Or the frequency range corresponding to the octave with the largest SEP statistic contribution is used as the fault audio characteristic of the power transformer, and is compared with an expert experience rule base to predict or obtain the fault type of the power transformer.
In the embodiment, step S1 takes an audio signal of the power transformer as a weak periodic or quasi-periodic disturbance signal of the chaotic system, and detects the audio signal of the power transformer through calculation and identification by utilizing the characteristic that the phase state of the chaotic system is changed essentially due to perturbation and sensitivity of the chaotic system to parameters. According to the nonlinear dynamics system, when the chaotic state and the large-scale periodic state are in the same state, the corresponding system phase state and the system state are different, and the transition from the chaotic phase state to the large-scale periodic phase state of the system is used as the detection basis of a weak periodic or quasi-periodic signal. The chaotic system is based on Duffing-Holmes, and the specific method comprises the following steps:
s101, the Duffing-Holmes oscillator equation is
Wherein k is the damping ratio, -x+x 3 For nonlinear restoring force term, rcos (ωt) is periodic strategy force or excitation signal, r is periodic strategy force amplitude, poincare mapping
Chaos in the sense of Smale horseshoes exists;
s102, calculating a threshold value for generating chaos by a Melnikov method, wherein R/k satisfies inequality R/k > R m
Wherein R is m For m harmonic rail bifurcation values of parameter R/k, when R/k crosses bifurcation value R m Chaos of timeThe topological nature of the vector field of the system in phase space will jump; when R/k is smaller than R m When the system is in a chaotic state, a periodic closed rail does not exist; when R/k is greater than R m When the system is in a large period state, a cluster of period closed tracks exist.
As shown in fig. 3, the specific calculation method of the mel frequency cepstrum technique in step S2 includes the following steps:
s201, pre-emphasis is carried out on the audio fragment by adopting a high-pass filter, so that the problem that the loss of a high-frequency component signal is larger than that of a low-frequency signal in the transmission process is solved:
H(z)＝1-bz -1
wherein b is 0.98;
s202, in order to avoid mutation at two end points during discrete FFT conversion, a Hamming window is adopted to carry out windowing treatment on the audio fragment:
w[n]＝0.54-0.46cos(2nπ/L)
wherein n represents a sampling point number, and L represents a window length;
s203, performing N-point discrete Fourier transform DFT on each frame of time domain signal, wherein N is greater than or equal to L:
s204, arranging a plurality of band-pass filters H in the sound frequency spectrum range m (k) M is more than or equal to 0 and less than M, wherein M is the number of filters;
the filter has triangular filtering characteristics, the center frequency is f (m), and the filter has equal bandwidth in the Mel frequency range; transforming the actual frequency domain into the Mel frequency domain by adopting the following method
m(f)＝1125ln(1+f/700)
Mapping the limited practical frequency range [ fl, fh ] to the Mel frequency range [ Fsell, fsell ] and equally dividing the Mel frequency range into M+1 parts to obtain M Mel center frequencies, and mapping the obtained M Mel center frequencies to the practical frequency domain to obtain the center frequencies of M triangular filters in the practical frequency domain;
according to the center frequency of the actual frequency domain, determining the transfer function of the Mel filter bank, wherein m is the filter sequence number:
the logarithmic energy of the frequency domain signal X (k) output by the mth Mel filter is calculated, and the expression is:
wherein E (m) is logarithmic energy, H m (k) For a mel filter bank, X (k) is the frequency domain signal;
s205, obtaining Mel cepstrum coefficient through discrete cosine transform DCT, wherein the expression is
Wherein: c (n) is the nth frequency cepstrum coefficient, E (M) is logarithmic energy, and M is the number of Mel filters, namely the output dimension.
In step S3 of this embodiment, principal component analysis is used to extract principal component of the Mel cepstrum feature of the audio signal, where X represents the Mel cepstrum feature matrix of the audio signal, R represents the correlation coefficient matrix of X, and the eigenvalue is calculated according to the eigenvalue |r- λe|=0, i.e. the eigenvalue is solved
r n λ m +r n-1 λ m-1 +…+r 1 λ+r 0 ＝0
Obtaining a characteristic value lambda 1 ,λ 2 ,…,λ m And arranging the characteristic values in order from large to small, i.e. lambda 1 ≥λ 2 ≥…≥λ m The method comprises the steps that (1) the principal component is determined according to a set cumulative contribution rate threshold value; wherein the contribution rate of the single feature isThe first k eigenvalues add up to contribute +.>Taking the first, second, … and p (p is less than or equal to m) main components corresponding to the characteristic values with the accumulated contribution reaching 85-95%.
Further, in step S4, the method for training the prediction model by using the vector machine algorithm includes:
given a training sample set of { (x) i ,y i ) I=1, 2, …, l }, where x i ∈R n Representing the input vector, y i E R represents the output result, and nonlinear mapping is adoptedInput vector x i Mapping to higher dimensional feature space R k (k>n) constructing an optimal hyperplane in the space>The "total deviation" of all sample points from the hyperplane is minimized,
wherein ω represents a weight coefficient vector; b is a bias constant;
using the epsilon-insensitive loss function, the deviation of the sample point from the optimal hyperplane can be expressed as
c(x,y,f(x))＝max(0,|y-f(x)|-ε)
Wherein ε represents the allowable error;
adding relaxation factor xi i 、When there is an error in the division, ζ i 、/>Are all greater than 0; when the division is error-free, xi i 、/>All take 0, convert to solve the optimization objective function minimization problem:
wherein, C represents penalty factor, and is converted into convex quadratic optimization problem; constraint optimization problem is solved by Lagrange coefficient method according to optimization condition KKT
Wherein alpha is i 、Representing Lagrange coefficients, SV representing the support vector.
In the step S4 of this embodiment, the quantum particle swarm algorithm is adopted to optimize the optimal super-parameters of the SVM, and a heuristic intelligent optimization algorithm simulating the behavior of the biological swarm or the law of biological evolution is used to search or approach the global optimal solution of the nonlinear complex space in a random or nearly random manner, so that the searching efficiency can be effectively improved. Taking a quantum particle swarm algorithm (Quantum Particle Swarm Optimization, QSPO) as an example, the attribute of the particle moving direction is eliminated, the randomness of the particle positions is increased, the parameters required to be set are less than those of the particle swarm algorithm, and the influence of the initial random positions of the particles on the search results is reduced. Let the particle population size be m, the particle dimension be D, the position of each particle represents the current solution of the particle, and the ith particle has the following properties:
current position: x is x i ＝(x i1 ,x i2 ,…,x iD )；
Historical optimal position: p is p i ＝(p i1 ,p i2 ,…p iD )；
The average value of the historic optimal positions of all particles is
The next position of the particle is updated as follows
Wherein g best Is a global optimal position; alpha is a compression-expansion factor;and u is the value of the uniform distribution on (0, 1); the probability of taking the sum is 0.5.
In this embodiment, the flow of the QSPO algorithm is shown in fig. 4, and the specific steps are as follows:
(1) Setting parameters such as population scale m, particle dimension D, compression-expansion factor alpha, maximum iteration number iter, particle solution space range and the like;
(2) Initializing the current position x of the particle i Particle historic optimal position p i Particle history fitness function value fit pi Global optimum position g best ；
(3) Calculating the current fitness function value fit of all particles i Searching the optimal fitness function value fit of all the current particles best And its corresponding optimal position;
(4) Updating global historical optimal position gb est And the corresponding historical optimal population fitness function value fit gbest Updating all particle historic optimal positions p i And its historical optimum population fitness fit pi ；
(5) Repeating the step (3-4) until reaching the maximum iteration number item, outputting the current global optimal position g best 。
As shown in fig. 5, the 1/3 octave power-spectrum-based method in step S5 calculates the power transformer pitchThe frequency signal and the power transformer audio signal are windowed and framed according to the method described in step S202, and the 1/3 octave amplitude calculation result of each frame is recorded as x= (x) 1 ,x 2 ,…,x 29 ) And calculates T of the power transformer audio signal based on 1/3 octave amplitude 2 And SEP statistics (also known as Q statistics),
T 2 statistics measure the change in principal component space of sample x:
wherein Λ= (λ) 1 ,λ 2 ,...,λ p )，Is a control limit with a confidence level of alpha;
the usual calculation method of the control limit is:
wherein F is α (p, n-p) is the F distribution value with p and n-p degrees of freedom with confidence α;
the SEP index measures the change in projection of the sample vector x in the residual space:
wherein,a control limit indicating a confidence level α;
the calculation formula of (2) is as follows:
wherein,i＝1,2,3，/> is the eigenvalue of the covariance matrix of X, C α A threshold value with the confidence degree alpha is distributed for standard normal;
the definition of the T2-based contribution graph is as follows:
the SPE-based contribution graph is defined as follows:
wherein,
and when the T2 or SEP statistic exceeds a threshold value, the frequency range corresponding to the octave with the largest contribution graph is regarded as the fault characteristic frequency, and the fault type possibly occurring in the power transformer is obtained by comparing the frequency range with expert experience rules.
Example 2:
as shown in fig. 1, the utility model also provides a power transformer fault prediction and diagnosis system based on audio characteristics, which comprises a field device layer 1, a sensing device layer 2, an edge gateway layer 3, an AI big data platform 4 and an external service layer 5; the sensing equipment layer 2 uploads the real-time collected power transformer audio data to the edge gateway layer 3; the edge gateway layer 3 predicts the running state of the power transformer according to the published prediction model and the received real-time audio data, and if the power transformer is in a fault state, the edge gateway layer 3 uploads the audio data of the power transformer to the AI big data platform 4 for fault diagnosis; the AI big data platform 4 generates an alarm event after receiving the fault information, and performs fault diagnosis on the power transformer according to the received real-time data; the external service layer 5 provides a presentation page of the power transformer fault prediction and diagnosis system.
The AI big data platform 4 of the embodiment can perform timing training and issue a prediction model, evaluate the health state of the power transformer at regular time, manage equipment, process data, serve data and control equipment emergently.
The sensing device layer 2 of the embodiment also uploads the power transformer temperature data acquired in real time and/or the infrared image to the edge gateway layer 3; fault detection and determination may also be made based on other measured data or infrared images of the power transformer.
The specific embodiments described herein are offered by way of example only to illustrate the spirit of the utility model. Those skilled in the art may make various modifications or additions to the described embodiments or substitutions thereof without departing from the spirit of the utility model or exceeding the scope of the utility model as defined in the accompanying claims.
Claims (9)
1. The power transformer fault prediction and diagnosis method based on the audio frequency characteristics is characterized by comprising the following specific steps:
s1, detecting effective signals of audio data of a power transformer under a noise background based on a chaotic oscillator;
s2, aiming at the audio effective signal extracted in the step S1, calculating a logarithmic energy spectrum on a nonlinear Mel scale by adopting a Mel frequency cepstrum technology, and taking the logarithmic energy spectrum as a characteristic quantity of the audio signal of the power transformer;
s3, calculating principal components of the characteristic quantity of the audio signal of the power transformer by adopting a principal component analysis method;
s4, adopting a quantum particle swarm optimization vector machine algorithm to perform super-parameter training on a power transformer fault prediction model;
s5, if the power transformer is in a fault state, extracting fault characteristic frequency range amplitude by adopting a 1/3 octave algorithm, and calculating T of an audio signal of the power transformer based on the 1/3 octave amplitude 2 And SEP statistics, find T 2 Or the SEP statistic exceeds a threshold value for T 2 Or the frequency range corresponding to the octave with the largest SEP statistic contribution graph is used as the fault audio characteristic of the power transformer, and is compared with an expert experience rule base to predict/obtain the fault type of the power transformer;
the 1/3 octave power spectrum-based method in the step S5 calculates the power transformer audio signal, the power transformer audio signal is windowed and framed according to the method in the step S202, and the 1/3 octave amplitude calculation result of each frame is recorded as x= (x) 1 ,x 2 ,…,x 29 ) And calculates T of the power transformer audio signal based on 1/3 octave amplitude 2 And SEP statistics, T 2 Statistics measure the change in principal component space of sample x:
wherein Λ= (λ) 1 ,λ 2 ,...,λ p )，Is a control limit with a confidence level of alpha;
the calculation method of the control limit comprises the following steps:
wherein F is α (p, n-p) is the F distribution value with p and n-p degrees of freedom with confidence α;
the SEP index measures the change in projection of the sample vector x in the residual space:
wherein,a control limit indicating a confidence level α;
the calculation formula of (2) is as follows:
wherein, is the eigenvalue of the covariance matrix of X, C α A threshold value with the confidence degree alpha is distributed for standard normal;
based on T 2 The definition of the contribution graph of (2) is:
SPE-based contribution graphs are defined as:
wherein,
when the T is detected 2 Or SEP systemAnd after the measurement exceeds the threshold value, the frequency range corresponding to the octave with the largest contribution graph is used as the fault characteristic frequency, and the fault type possibly occurring in the power transformer is obtained by comparing the fault characteristic frequency with expert experience rules.
2. The method for predicting and diagnosing a power transformer fault based on audio features as claimed in claim 1, wherein the step S1 uses the power transformer audio signal as a weak periodic or quasi-periodic disturbance signal of the chaotic system, the power transformer audio signal is detected through calculation and identification, and the transition of the system from the chaotic phase state to the large-scale periodic phase state is used as the detection basis of the weak periodic or quasi-periodic signal, and the chaotic system is based on Duffing-Holmes, and the specific method comprises the following steps:
s101, the Duffing-Holmes oscillator equation is
Wherein k is the damping ratio, -x+x 3 R cos (ωt) is periodic strategy power or excitation signal, r is periodic strategy power amplitude, and Poincare maps chaos in the meaning of Smale horseshoe;
s102, calculating a threshold value for generating chaos by a Melnikov method, wherein R/k satisfies inequality R/k > R m ；
Wherein R is m For m harmonic rail bifurcation values of parameter R/k, when R/k crosses bifurcation value R m When the chaotic system jumps in the topological property of the vector field in the phase space; when R/k is smaller than R m When the system is in a chaotic state, a periodic closed rail does not exist; when R/k is greater than R m When the system is in a large period state, a cluster of period closed tracks exist.
3. The power transformer fault prediction and diagnosis method based on audio features according to claim 1, wherein the mel frequency cepstrum technique specific calculation method of step S2 comprises the following steps:
s201, pre-emphasis is carried out on the audio fragment by adopting a high-pass filter, so that the problem that the loss of a high-frequency component signal is larger than that of a low-frequency signal in the transmission process is solved:
H(z)＝1-bz -1
wherein b is 0.98;
s202, in order to avoid mutation at two end points during discrete FFT conversion, a Hamming window is adopted to carry out windowing treatment on the audio fragment:
w[n]＝0.54-0.46cos(2nπ/L)
wherein n represents a sampling point number, and L represents a window length;
s203, performing N-point discrete Fourier transform DFT on each frame of time domain signal, wherein N is greater than or equal to L:
s204, arranging a plurality of band-pass filters H in the sound frequency spectrum range m (k) M is more than or equal to 0 and less than M, wherein M is the number of filters;
the filter has triangular filtering characteristics, the center frequency is f (m), and the filter has equal bandwidth in a Mel frequency range; transforming the actual frequency domain into the mel frequency domain
m(f)＝1125ln(1+f/700)
Mapping the limited actual frequency range [ fl, fh ] to the Mel frequency range [ Fsell, fsell ] and equally dividing the Mel frequency range into M+1 parts to obtain M Mel center frequencies, and mapping the M Mel center frequencies to an actual frequency domain to obtain the center frequencies of M triangular filters in the actual frequency domain;
according to the center frequency of the actual frequency domain, determining the transfer function of the Mel filter bank, wherein m is the filter sequence number:
the logarithmic energy of the frequency domain signal X (k) output by the mth Mel filter is calculated, and the expression is:
wherein E (m) is logarithmic energy, H m (k) For a mel filter bank, X (k) is the frequency domain signal;
s205, obtaining the mel cepstrum coefficient through discrete cosine transform DCT, wherein the expression is
Wherein: c (n) is the nth frequency cepstrum coefficient, E (M) is logarithmic energy, and M is the number of Mel filters, namely the output dimension.
4. The method for predicting and diagnosing faults of a power transformer based on audio features as claimed in claim 1, wherein the step S3 extracts principal components of the audio mel-frequency cepstrum features by principal component analysis method, uses X to represent mel-frequency cepstrum feature matrix of audio signal, R to represent correlation coefficient matrix of X, calculates eigenvalues according to the eigenvalue |r- λe|=0, i.e. solves the eigenvalue
r n λ m +r n-1 λ m-1 +…+r 1 λ+r 0 ＝0
Obtaining a characteristic value lambda 1 ,λ 2 ,…,λ m And arranging the characteristic values in order from large to small, i.e. lambda 1 ≥λ 2 ≥…≥λ m The method comprises the steps that (1) the principal component is determined according to a set cumulative contribution rate threshold value; wherein the contribution rate of the single feature isThe first k eigenvalues add up to contribute +.>Taking the first, second and … corresponding to the characteristic value with the accumulated contribution reaching 85-95 percent,The p (p.ltoreq.m) th principal component.
5. The method for predicting and diagnosing a fault of a power transformer based on audio features as claimed in claim 1, wherein the method for training the prediction model by using a support vector machine algorithm in step S4 is as follows:
given a training sample set of { (x) i ,y i ) I=1, 2, …, l }, where x i ∈R n Representing the input vector, y i E R represents the output result, and nonlinear mapping is adoptedInput vector x i Mapping to higher dimensional feature space R k (k>n) constructing an optimal hyperplane in space>The "total deviation" of all sample points from the hyperplane is minimized,
wherein ω represents a weight coefficient vector; b is a bias constant;
using the epsilon-insensitive loss function, the deviation of the sample point from the optimal hyperplane can be expressed as
c(x,y,f(x))＝max(0,|y-f(x)|-ε)
Wherein ε represents the allowable error;
adding relaxation factor xi i 、When there is an error in the division, ζ i 、/>Are all greater than 0; when the division is error-free, xi i 、/>All take 0, convert to solve the optimization objective function minimization problem:
wherein, C represents penalty factor, and is converted into convex quadratic optimization problem; adopting Lagrange coefficient method and solving constraint optimization problem according to optimization condition KKT:
wherein alpha is i 、Representing Lagrange coefficients, SV representing the support vector.
6. The method for predicting and diagnosing faults of a power transformer based on audio features as claimed in claim 1, wherein the super-parameters optimized by adopting the quantum particle swarm algorithm in the step S4 are as follows: let the particle population scale be m, the particle dimension be D, the position of each particle represents the current solution of the particle, the current position and the historical optimal position of the ith particle are:
current position: x is x i ＝(x i1 ,x i2 ,…,x iD )；
Historical optimal position: p is p i ＝(p i1 ,p i2 ,…p iD )；
The average value of the historical optimal positions of the particles is as follows:
the next position update of the particle is:
wherein g best Is a global optimal position; alpha is a compression-expansion factor;and u is the value of the uniform distribution on (0, 1); the probability of taking the sum is 0.5.
7. An audio feature-based power transformer fault prediction and diagnosis system using the audio feature-based power transformer fault prediction and diagnosis method according to any one of claims 1 to 6, characterized by comprising a field device layer (1), a sensing device layer (2), an edge gateway layer (3), an AI big data platform (4), an external service layer (5); the sensing equipment layer (2) uploads the real-time collected power transformer audio data to the edge gateway layer (3); the edge gateway layer (3) predicts the running state of the power transformer according to the published prediction model and the received real-time audio data, and if the power transformer is in a fault state, the edge gateway layer (3) uploads the audio data of the power transformer to the AI big data platform (4) for fault diagnosis; the AI big data platform (4) generates an alarm event after receiving the fault information, and performs fault diagnosis on the power transformer according to the received real-time data; the external service layer (5) provides a display page of the power transformer fault prediction and diagnosis system.
8. The audio feature-based power transformer fault prediction and diagnosis system according to claim 7, wherein the AI big data platform (4) can perform timing training and release prediction model, evaluate power transformer health status at timing, manage equipment, process data, service data, and control equipment urgently.
9. The power transformer fault prediction and diagnosis system based on audio features according to claim 7, characterized in that the device sensing layer (2) also uploads power transformer temperature data collected in real time and/or infrared images to the edge gateway layer (3).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011187629.3A CN112395959B (en) | 2020-10-30 | 2020-10-30 | Power transformer fault prediction and diagnosis method and system based on audio frequency characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011187629.3A CN112395959B (en) | 2020-10-30 | 2020-10-30 | Power transformer fault prediction and diagnosis method and system based on audio frequency characteristics |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112395959A CN112395959A (en) | 2021-02-23 |
| CN112395959B true CN112395959B (en) | 2024-01-02 |
Family
ID=74598465
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011187629.3A Active CN112395959B (en) | 2020-10-30 | 2020-10-30 | Power transformer fault prediction and diagnosis method and system based on audio frequency characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112395959B (en) |
Families Citing this family (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112880750A (en) * | 2021-03-26 | 2021-06-01 | 山东和兑智能科技有限公司 | Transformer multidimensional comprehensive online monitoring intelligent diagnosis system |
| CN113076860B (en) * | 2021-03-30 | 2022-02-25 | 南京大学环境规划设计研究院集团股份公司 | Bird detection system under field scene |
| CN113075512A (en) * | 2021-04-01 | 2021-07-06 | 陕西省地方电力（集团）有限公司 | Sound detection-based transformer discharge fault diagnosis method and system |
| CN113253156B (en) * | 2021-05-17 | 2023-01-06 | 国网江苏省电力有限公司检修分公司 | Sound monitoring-based latent defect diagnosis method for transformer |
| CN113608958B (en) * | 2021-07-06 | 2023-05-23 | 国网宁夏电力有限公司电力科学研究院 | Power equipment online monitoring error data diagnosis method based on time sequence chaos characteristic |
| 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 |
| CN113740779B (en) * | 2021-09-10 | 2023-09-19 | 杭州电力设备制造有限公司 | Distribution transformer state discrimination method, system, device and storage medium |
| CN114019310B (en) * | 2021-11-08 | 2024-02-20 | 国网湖南省电力有限公司 | Open-circuit fault pre-judging method, equipment and medium for energy storage converter |
| CN114137444B (en) * | 2021-11-29 | 2024-04-02 | 国网山东省电力公司日照供电公司 | Transformer running state monitoring method and system based on acoustic signals |
| CN114279727B (en) * | 2021-12-03 | 2024-02-09 | 国网江苏省电力有限公司宿迁供电分公司 | Method for judging aging of transformer cooling device |
| CN114238368A (en) * | 2021-12-07 | 2022-03-25 | 广东电网有限责任公司广州供电局 | Transformer fault diagnosis method and device, computer equipment and readable storage medium |
| CN115860714B (en) * | 2022-11-28 | 2023-08-08 | 珠海德瑞斯科技有限公司 | Power equipment safe operation management system and method based on industrial Internet |
| CN116007742A (en) * | 2022-11-30 | 2023-04-25 | 中国南方电网有限责任公司超高压输电公司广州局 | Transformer abnormal sound detection method, device, equipment and storage medium |
| CN117150791B (en) * | 2023-09-05 | 2024-03-26 | 正泰电气股份有限公司 | Method and device for optimally designing main insulation structure of transformer |
| CN117109730B (en) * | 2023-10-23 | 2024-01-12 | 山东和兑智能科技有限公司 | Transformer fault diagnosis system and method based on artificial intelligence |
| CN117131366B (en) * | 2023-10-26 | 2024-02-06 | 北京国电通网络技术有限公司 | Transformer maintenance equipment control method and device, electronic equipment and readable medium |
| CN117713385B (en) * | 2024-02-05 | 2024-05-03 | 威海双城电气有限公司 | Electric power detection system and method based on combination of quantum computation and artificial intelligence |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103884421A (en) * | 2014-03-24 | 2014-06-25 | 重庆邮电大学 | Duffing oscillator weak-signal detection method based on united denoising and pseudo Hamiltonian |
| CN108535648A (en) * | 2018-03-29 | 2018-09-14 | 上海交通大学 | Method of Motor Fault Diagnosis and system |
| CN108562853A (en) * | 2018-03-29 | 2018-09-21 | 上海交通大学 | Method of Motor Fault Diagnosis based on error correcting output codes support vector machines and system |
| CN109357749A (en) * | 2018-09-04 | 2019-02-19 | 南京理工大学 | A kind of power equipment audio signal analysis method based on DNN algorithm |
| CN110208022A (en) * | 2019-06-12 | 2019-09-06 | 济南雷森科技有限公司 | Power equipment multiple features audio-frequency fingerprint fault diagnosis method and system based on machine learning |
| CN110335617A (en) * | 2019-05-24 | 2019-10-15 | 国网新疆电力有限公司乌鲁木齐供电公司 | A kind of noise analysis method in substation |
| 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 |
| CN110969096A (en) * | 2019-11-07 | 2020-04-07 | 国电南京自动化股份有限公司 | Motor fault mode diagnosis method based on particle swarm optimization support vector machine |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8055401B2 (en) * | 2009-01-27 | 2011-11-08 | GM Global Technology Operations LLC | Transmission auxiliary pump diagnosis method and apparatus |
-
2020
- 2020-10-30 CN CN202011187629.3A patent/CN112395959B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103884421A (en) * | 2014-03-24 | 2014-06-25 | 重庆邮电大学 | Duffing oscillator weak-signal detection method based on united denoising and pseudo Hamiltonian |
| CN108535648A (en) * | 2018-03-29 | 2018-09-14 | 上海交通大学 | Method of Motor Fault Diagnosis and system |
| CN108562853A (en) * | 2018-03-29 | 2018-09-21 | 上海交通大学 | Method of Motor Fault Diagnosis based on error correcting output codes support vector machines and system |
| CN109357749A (en) * | 2018-09-04 | 2019-02-19 | 南京理工大学 | A kind of power equipment audio signal analysis method based on DNN algorithm |
| CN110335617A (en) * | 2019-05-24 | 2019-10-15 | 国网新疆电力有限公司乌鲁木齐供电公司 | A kind of noise analysis method in substation |
| CN110208022A (en) * | 2019-06-12 | 2019-09-06 | 济南雷森科技有限公司 | Power equipment multiple features audio-frequency fingerprint fault diagnosis method and system based on machine learning |
| 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 |
| CN110969096A (en) * | 2019-11-07 | 2020-04-07 | 国电南京自动化股份有限公司 | Motor fault mode diagnosis method based on particle swarm optimization support vector machine |
Non-Patent Citations (4)
| Title |
|---|
| 基于DUFFING阵子的接地网故障诊断弱信号幅值检测新方法;倪云峰;康海雷;刘健;王森;;电测与仪表(10);全文 * |
| 基于梅尔频率倒谱系数与翻转梅尔频率倒谱系数的说话人识别方法;胡峰松;张璇;;计算机应用(09);全文 * |
| 小型电机生产过程在线品质检测方法研究;罗炳聪;樊可清;陈英武;;测控技术(12);全文 * |
| 电力变压器故障自动化实时检测技术;张瑞芯;;机电信息(20);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112395959A (en) | 2021-02-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112395959B (en) | Power transformer fault prediction and diagnosis method and system based on audio frequency characteristics | |
| WO2021232655A1 (en) | Method for evaluating mechanical state of high-voltage shunt reactor on the basis of vibration feature | |
| CN111325095B (en) | Intelligent detection method and system for equipment health state based on acoustic wave signals | |
| US9347432B2 (en) | System and method for enhanced operation of wind parks | |
| Soualhi et al. | Prognosis of bearing failures using hidden Markov models and the adaptive neuro-fuzzy inference system | |
| CN208834172U (en) | The device and processor platform of situation for monitoring of structures | |
| CN112201260B (en) | Transformer running state online detection method based on voiceprint recognition | |
| CN103810374A (en) | Machine fault prediction method based on MFCC feature extraction | |
| CN105071983A (en) | Abnormal load detection method for cloud calculation on-line business | |
| Liu et al. | Combined forecasting method of dissolved gases concentration and its application in condition-based maintenance | |
| CN112115643B (en) | Smart train service life non-invasive prediction method | |
| CN116380445B (en) | Equipment state diagnosis method and related device based on vibration waveform | |
| CN110412368A (en) | Electrical equipment online supervision method and system based on Application on Voiceprint Recognition | |
| Shan et al. | Modeling of temperature effect on modal frequency of concrete beam based on field monitoring data | |
| CN117371207A (en) | Extra-high voltage converter valve state evaluation method, medium and system | |
| Hu et al. | A hybrid model for predicting the degradation trend of hydropower units based on deep learning | |
| CN117076955A (en) | Fault detection method and system for high-voltage frequency converter | |
| CN117217730A (en) | Power equipment fault identification method, device, equipment, medium and product | |
| CN116796189A (en) | Aerosol extinction coefficient profile prediction method based on deep learning technology | |
| CN115130788B (en) | Big data analysis-based power utilization prediction system | |
| Sun et al. | Jointed task of multi-scale CNN based power transformer fault diagnosis with vibration and sound signals | |
| CN115406522A (en) | Power plant equipment running state research and application based on voiceprint recognition | |
| WO2022141623A1 (en) | Train compartment vibration monitoring method, vibration signal feature library establishment method and vibration signal feature library application method | |
| CN115455684A (en) | Oil-immersed transformer non-contact state evaluation method and system based on noise analysis | |
| Yu et al. | Research and application of intelligent diagnosis method of mechanical fault based on transformer vibration and noise and BP neural network |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20230614 Address after: 213031 Tianhe PV Industrial Park No. 2, Xinbei District, Changzhou, Jiangsu Applicant after: TRINASOLAR Co.,Ltd. Address before: Room 2206, building 1, 650 Bin'an Road, Changhe street, Binjiang District, Hangzhou City, Zhejiang Province 310051 Applicant before: Tianheyun Energy Internet Technology (Hangzhou) Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |