CN114121025A - Voiceprint fault intelligent detection method and device for substation equipment - Google Patents

Voiceprint fault intelligent detection method and device for substation equipment Download PDF

Info

Publication number
CN114121025A
CN114121025A CN202111404981.2A CN202111404981A CN114121025A CN 114121025 A CN114121025 A CN 114121025A CN 202111404981 A CN202111404981 A CN 202111404981A CN 114121025 A CN114121025 A CN 114121025A
Authority
CN
China
Prior art keywords
voiceprint
detected
signal
short
time
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
Application number
CN202111404981.2A
Other languages
Chinese (zh)
Inventor
卢大玮
邱镇
廖逍
徐海青
孙飞
白景坡
王兴涛
崔蔚
黄晓光
王维佳
梁翀
李小宁
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
State Grid Information and Telecommunication Co Ltd
Anhui Jiyuan Software Co Ltd
Original Assignee
State Grid Information and Telecommunication Co Ltd
Anhui Jiyuan Software Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by State Grid Information and Telecommunication Co Ltd, Anhui Jiyuan Software Co Ltd filed Critical State Grid Information and Telecommunication Co Ltd
Priority to CN202111404981.2A priority Critical patent/CN114121025A/en
Publication of CN114121025A publication Critical patent/CN114121025A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L17/00—Speaker identification or verification techniques
    • G10L17/26—Recognition of special voice characteristics, e.g. for use in lie detectors; Recognition of animal voices
    • G—PHYSICS
    • 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
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C1/00—Registering, indicating or recording the time of events or elapsed time, e.g. time-recorders for work people
    • G07C1/20—Checking timed patrols, e.g. of watchman
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L17/00—Speaker identification or verification techniques
    • G10L17/04—Training, enrolment or model building
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
    • G10L25/24—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being the cepstrum
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/27—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique
    • G10L25/30—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique using neural networks
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J13/00—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network
    • H02J13/00001—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by the display of information or by user interaction, e.g. supervisory control and data acquisition systems [SCADA] or graphical user interfaces [GUI]
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J13/00—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network
    • H02J13/00002—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by monitoring
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/16—Electric power substations

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Human Computer Interaction (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Evolutionary Computation (AREA)
  • Acoustics & Sound (AREA)
  • Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computational Linguistics (AREA)
  • Signal Processing (AREA)
  • Power Engineering (AREA)
  • Probability & Statistics with Applications (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

After a voiceprint signal to be detected is obtained, voiceprint feature extraction is carried out on the voiceprint signal to be detected, and a feature vector corresponding to the voiceprint signal to be detected is obtained; inputting the characteristic vector serving as an input parameter into a preset voiceprint recognition model to obtain a working condition label corresponding to the voiceprint signal to be detected, carrying out model training based on transfer learning by the preset voiceprint recognition model through a voiceprint feature analysis network, and establishing voiceprint recognition models of the substation equipment under different working conditions; and performing similarity comparison on the working condition label and the target template output by utilizing a cosine similarity measurement algorithm formula so as to realize intelligent voiceprint fault detection of the substation equipment. By the aid of the voice print fault detection method and device, voice print fault detection efficiency of the substation equipment can be effectively improved, labor inspection cost is reduced, and therefore the intelligent level of inspection of the substation equipment is comprehensively improved.

Description

Voiceprint fault intelligent detection method and device for substation equipment
Technical Field
The application relates to the technical field of intelligent detection, in particular to a voiceprint fault intelligent detection method and device for substation equipment.
Background
The transformer substation equipment inspection is basic work for ensuring safe operation of a transformer substation and improving power supply reliability, and the transformer substation mainly inspects operation sound states of transformer equipment such as a transformer, a circuit breaker, a disconnecting switch, a mutual inductor, a lightning arrester, a compensating device and a bus according to relevant transformer substation inspection standards and standards.
The common manual inspection method judges whether the vibration sound of the equipment is normal or not by listening to the sound emitted by the equipment and by inspection personnel and experience, lacks an intelligent monitoring mode for the operation sound of the power transformation equipment, and cannot quickly and effectively monitor whether a fault occurs or not for sound data (including voiceprint information such as an operation state, a fault reason and the like) generated in the operation process of the power transformation equipment.
Disclosure of Invention
In view of the problems in the foregoing, the application provides a voiceprint fault intelligent detection method and device for substation equipment, which are used for effectively improving voiceprint fault detection efficiency of the substation equipment and reducing labor inspection cost, so that the intelligent level of inspection of the substation equipment is comprehensively improved.
In order to achieve the above object, the present application provides the following technical solutions:
a voiceprint fault intelligent detection method for substation equipment comprises the following steps:
acquiring a voiceprint signal to be detected, and extracting voiceprint characteristics of the voiceprint signal to be detected to obtain a characteristic vector corresponding to the voiceprint signal to be detected;
inputting the characteristic vector serving as an input parameter into a preset voiceprint recognition model to obtain a working condition label corresponding to the voiceprint signal to be detected, carrying out model training based on transfer learning by the preset voiceprint recognition model through a voiceprint feature analysis network, and establishing voiceprint recognition models of the substation equipment under different working conditions;
and performing similarity comparison on the working condition label and the target template output by utilizing a cosine similarity measurement algorithm formula so as to realize intelligent voiceprint fault detection of the substation equipment.
Further, the acquiring a voiceprint signal to be detected, and performing voiceprint feature extraction on the voiceprint signal to be detected to obtain a feature vector corresponding to the voiceprint signal to be detected includes:
acquiring a voiceprint signal to be detected, and preprocessing the voiceprint signal to be detected to obtain a plurality of framing signals;
respectively extracting short-time domain features and short-time frequency domain features from each frame signal, and generating an MFCC feature vector group, wherein the short-time domain features comprise short-time energy, a short-time average zero crossing rate and a short-time energy entropy, and the short-time frequency domain features are Mel frequency cepstrum coefficients;
and performing characteristic connection on the MFCC characteristic vectors in the MFCC characteristic vector group to obtain a characteristic vector corresponding to the voiceprint signal to be detected.
Further, before the feature concatenating the MFCC feature vectors in the MFCC feature vector group, the method further includes:
and introducing a weighting optimization method based on an F ratio into the MFCC feature vector group, and performing dimension reduction processing on the MFCC feature vectors in the MFCC feature vector group based on multi-modal features of PCA.
Further, the preprocessing the voiceprint signal to be detected to obtain a plurality of framing signals includes:
dividing the voiceprint signal to be detected collected in a preset time period into a plurality of short-time signals;
and after windowing each short-time signal by adopting a Hamming window, carrying out Fourier transform on each short-time signal to obtain a plurality of framing signals.
The utility model provides a voiceprint fault intelligent detection device towards substation equipment, includes:
the first processing unit is used for acquiring a voiceprint signal to be detected and extracting voiceprint characteristics of the voiceprint signal to be detected to obtain a characteristic vector corresponding to the voiceprint signal to be detected;
the second processing unit is used for inputting the characteristic vectors serving as input parameters into a preset voiceprint recognition model to obtain a working condition label corresponding to the voiceprint signal to be detected, and the preset voiceprint recognition model is used for carrying out model training based on transfer learning through a voiceprint feature analysis network to establish voiceprint recognition models of the substation equipment under different working conditions;
and the third processing unit is used for comparing the similarity of the working condition label and the target template output by utilizing a cosine similarity measurement algorithm formula so as to realize the intelligent voiceprint fault detection of the substation equipment.
Further, the first processing unit is configured to:
acquiring a voiceprint signal to be detected, and preprocessing the voiceprint signal to be detected to obtain a plurality of framing signals;
respectively extracting short-time domain features and short-time frequency domain features from each frame signal, and generating an MFCC feature vector group, wherein the short-time domain features comprise short-time energy, a short-time average zero crossing rate and a short-time energy entropy, and the short-time frequency domain features are Mel frequency cepstrum coefficients;
and performing characteristic connection on the MFCC characteristic vectors in the MFCC characteristic vector group to obtain a characteristic vector corresponding to the voiceprint signal to be detected.
Further, the first processing unit is further configured to:
and introducing a weighting optimization method based on an F ratio into the MFCC feature vector group, and performing dimension reduction processing on the MFCC feature vectors in the MFCC feature vector group based on multi-modal features of PCA.
Further, the first processing unit is further configured to:
dividing the voiceprint signal to be detected collected in a preset time period into a plurality of short-time signals;
and after windowing each short-time signal by adopting a Hamming window, carrying out Fourier transform on each short-time signal to obtain a plurality of framing signals.
A storage medium comprising a stored program, wherein when the program runs, a device where the storage medium is located is controlled to execute the voiceprint fault intelligent detection method for substation equipment.
An electronic device comprising at least one processor, and at least one memory, bus connected with the processor; the processor and the memory complete mutual communication through the bus; the processor is used for calling the program instructions in the memory so as to execute the voiceprint fault intelligent detection method for the substation equipment.
According to the method and the device for intelligently detecting the voiceprint faults facing the substation equipment, after the voiceprint signals to be detected are obtained, voiceprint feature extraction is carried out on the voiceprint signals to be detected, and feature vectors corresponding to the voiceprint signals to be detected are obtained; inputting the characteristic vector serving as an input parameter into a preset voiceprint recognition model to obtain a working condition label corresponding to the voiceprint signal to be detected, carrying out model training based on transfer learning by the preset voiceprint recognition model through a voiceprint feature analysis network, and establishing voiceprint recognition models of the substation equipment under different working conditions; and performing similarity comparison on the working condition label and the target template output by utilizing a cosine similarity measurement algorithm formula so as to realize intelligent voiceprint fault detection of the substation equipment. By the aid of the voice print fault detection method and device, voice print fault detection efficiency of the substation equipment can be effectively improved, labor inspection cost is reduced, and therefore the intelligent level of inspection of the substation equipment is comprehensively improved.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of an intelligent voiceprint fault detection method for substation equipment disclosed in an embodiment of the present application;
fig. 2 is a flowchart of a voiceprint fault detection method for a power transformation device disclosed in an embodiment of the present application;
fig. 3 is a schematic flow chart illustrating a process of obtaining a voiceprint signal to be detected and extracting voiceprint features of the voiceprint signal to be detected to obtain a feature vector corresponding to the voiceprint signal to be detected, disclosed in the embodiment of the present application;
fig. 4 is a schematic view of a process of extracting voiceprint signal characteristics of a power transformation device disclosed in the embodiment of the present application;
FIG. 5 is a schematic diagram of an MFCC feature extraction process disclosed in an embodiment of the present application;
fig. 6 is a diagram of a network structure for analyzing voiceprint characteristics disclosed in an embodiment of the present application;
FIG. 7 is a schematic diagram of parameter migration training disclosed in an embodiment of the present application;
fig. 8 is a schematic structural diagram of an intelligent voiceprint fault detection device for substation equipment disclosed in an embodiment of the present application;
fig. 9 is a schematic structural diagram of an electronic device disclosed in an embodiment of the present application.
Detailed Description
The operation and maintenance work of the transformer substation is an important guarantee for the safe and stable operation of the transformer substation, along with the continuous promotion of the construction of a large power grid, the number of equipment in the transformer substation is rapidly increased, and the operation and maintenance work of the equipment faces the outstanding contradiction of the rapid increase of the workload and the relative shortage of personnel. In recent years, computer technology and artificial intelligence technology are rapidly developed, so that the construction of an intelligent substation becomes an important measure for solving the problem of insufficient staff. The voiceprint recognition achieves the purpose of sound identification through characteristic analysis of various voice signals, namely, the theoretical basis of the voiceprint recognition is that each sound has unique characteristics, and different sounds can be effectively distinguished through the characteristics.
At present, voiceprint faults of substation equipment are mainly judged by manual work, although some monitoring sensing equipment are deployed in a substation, abnormal judgment is carried out on equipment operation sound by inspection personnel experience, and the quality, the efficiency and the real-time performance of manual evaluation are difficult to guarantee. The intelligent identification of the sound state attribute of the power transformation equipment is still blank, and due to the influences of various factors such as climate environment, noise interference, equipment models and acquisition devices, the research and application of the equipment voiceprint fault identification technology have many difficulties, and further deep research is still needed.
Therefore, for realizing intelligent recognition of voiceprint faults of the substation equipment, starting from the improvement of the accuracy of fault diagnosis and analysis, the application provides the voiceprint fault intelligent detection method for the substation equipment, which can accurately judge whether the voiceprint faults occur to the substation equipment and identify the types of the faults, effectively improves the voiceprint fault detection efficiency of the substation equipment, reduces the manual inspection cost, and therefore the intelligent level of inspection of the substation equipment is comprehensively improved.
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
Referring to fig. 1, a schematic flow chart of a voiceprint fault intelligent detection method for substation equipment provided in an embodiment of the present application is shown. As shown in fig. 1, an embodiment of the present application provides an intelligent voiceprint fault detection method for substation equipment, including the following steps:
s101: acquiring a voiceprint signal to be detected, and performing voiceprint feature extraction on the voiceprint signal to be detected to obtain a feature vector corresponding to the voiceprint signal to be detected.
It should be noted that voiceprint signals of the power transformation equipment are tedious and disordered, and the voiceprints of the power transformation equipment under different working conditions have high similarity in time domain and frequency domain, so that it is very difficult to directly analyze and identify the voiceprints. Therefore, by extracting features in the voiceprint signal, subsequent voiceprint analysis can be facilitated, and specifically, as shown in fig. 2, a voiceprint fault detection flow chart of the power transformation equipment is provided.
In this embodiment, as shown in fig. 3 and 4, the obtaining a voiceprint signal to be detected, and extracting voiceprint features of the voiceprint signal to be detected to obtain a feature vector corresponding to the voiceprint signal to be detected includes:
s301: the method comprises the steps of obtaining a voiceprint signal to be detected, preprocessing the voiceprint signal to be detected, and obtaining a plurality of framing signals.
In this embodiment of the application, the preprocessing the voiceprint signal to be detected to obtain a plurality of framing signals includes: dividing the voiceprint signal to be detected collected in a preset time period into a plurality of short-time signals; and after windowing each short-time signal by adopting a Hamming window, carrying out Fourier transform on each short-time signal to obtain a plurality of framing signals.
It should be noted that, the voiceprint signals of the power transformation equipment need to be preprocessed first, and the voiceprint signals collected in a time period are divided into a plurality of short-time signals to be processed, where the signal preprocessing includes two steps of framing and windowing.
In order to ensure the continuity of the signals of two adjacent frames, there is generally an overlap between the two frames. If the frame length is too short, the characteristic vector scale is smaller, and the representativeness is poor; if the frame length is too long, the voice signal changes too much, which affects the accuracy of the feature vector. Compared with a voice signal, a voiceprint signal of the power transformation equipment is stable, the frame length can be increased to obtain higher accuracy, and the recognition efficiency is seriously influenced by the overlong frame length. The voiceprint signal of the power transformation equipment adopts the frame length of 500 ms. In addition, because the voiceprint signals of the transformer equipment under the same working condition are stable, the continuity of two adjacent frames after framing is good, and the overlapping rate is 40% for reducing the calculated amount.
In the embodiment of the present application, in order to extract the framing features, discrete fourier transform needs to be performed on the framing features. However, the direct discrete fourier transform of the framing signal generates a large distortion, and therefore, each framing type is firstly windowed and then fourier transformed to increase continuity at both ends of the signal, thereby reducing the distortion caused by the fourier transform. It should be noted that, since the hamming window function has a better low-pass capability and can better reflect the frequency characteristics of the short-time signal, the hamming window is used in the embodiment of the present application to perform windowing on the framing signal.
S302: and respectively extracting short-time frequency domain features and short-time frequency domain features from each frame signal, and generating an MFCC feature vector group.
The short-time frequency domain features comprise short-time energy, a short-time average zero crossing rate and a short-time energy entropy, and the short-time frequency domain features are Mel frequency cepstrum coefficients; it should be noted that the short-time energy is energy for calculating noise signals of each sub-frame, and a vector formed by the short-time energy of all frames is a short-time energy feature; the short-time average zero-crossing rate is the frequency of signal waveform passing through a transverse axis in each sub-frame, and a vector formed by the short-time zero-crossing rates of all sub-frames is the short-time average zero-crossing rate characteristic; the short-time energy entropy is used for calculating the uniformity degree of energy distribution in each sub-frame, and vectors formed by the short-time energy entropies of all sub-frames are short-time energy entropy characteristics; the solving process of Mel frequency cepstrum coefficient MFCC includes five processes of Fast Fourier Transform (FFT), modulus, Mel filtering, logarithmic transform and Discrete Cosine Transform (DCT), as shown in FIG. 5. And respectively solving an MFCC feature vector for each preprocessed sub-frame to form a feature vector group together. Specifically, each frame signal is first FFT transformed and its modulus value taken, and then transformed to the Mel frequency domain by p Mel filter banks. The resulting set of p coefficients c (i) constitutes the MFCC feature vector for the frame.
S303: and performing characteristic connection on the MFCC characteristic vectors in the MFCC characteristic vector group to obtain a characteristic vector corresponding to the voiceprint signal to be detected.
In the embodiment of the application, the contribution rates of the short-time domain feature and the frequency domain feature to voiceprint recognition are different, and if the feature vector is directly used for subsequent analysis, the calculation amount is large, and the recognition rate can be reduced. Therefore, a weighting optimization method based on an F ratio is introduced into the MFCC feature vector group, and the purpose of improving the identification rate of the noise signals of the transformer is achieved by improving the F ratio of the feature vectors; then, the dimensionality of the feature vector formed by the connection of the short-time domain and the frequency domain is reduced through PCA (principal component analysis), and the dimensionality of the feature vector is reduced as much as possible on the premise of ensuring that information is not lost, so that the calculation complexity is reduced.
Before the feature concatenation of the MFCC feature vectors in the MFCC feature vector group, the method further includes: and introducing a weighting optimization method based on an F ratio into the MFCC feature vector group, and performing dimension reduction processing on the MFCC feature vectors in the MFCC feature vector group based on multi-modal features of PCA.
The MFCC feature vector weighting optimization based on the F ratio is as follows: the F ratio is a commonly used index in the field of measurement statistics and is used for inspecting the effectiveness of each dimension component in the multidimensional data. The component with high F ratio shows that the component has smaller variance in the same type of data and larger variance in different types of data, and has higher efficiency when used for data classification. The component F ratio F (k) of the k-th dimension of the feature vector is calculated as follows:
Figure BDA0003371957350000091
wherein N is the number of identification objects, uk(i) Is the k-dimension component of the ith object feature vector; u. ofkIs an average value; x is the number ofk(i) Identifying the k-dimension component of all samples of the object for the i-th dimension; n isiThe number of samples of the i-th identification object.
It should be further noted that if the F ratio is greater than 1, the component can function to distinguish different types of noise signals, and the larger the F ratio, the stronger the distinguishing capability. And taking the F ratio as a weight, carrying out weighting processing on the output feature vector, enhancing components which contribute more to identification, and inhibiting components which contribute less, thereby improving the accuracy of noise identification.
The PCA-based multi-modal feature dimensionality reduction: the basic idea of PCA is to linearly combine the components of the original eigenvectors in each dimension to generate a new eigenvector, so that the components of the new eigenvector in each dimension are uncorrelated with each other and the variance is as large as possible. And (3) setting the dimension of a combined feature vector formed by the voiceprint time domain feature vectors and the frequency domain feature vectors as m, and obtaining fn feature vectors in total according to preprocessing. From this a matrix X of m X fn is constructed. And (3) solving a correlation matrix R of X, and simultaneously solving the eigenvalue and the eigenvector of R:
Figure BDA0003371957350000092
R＝XTX/(n-1)
calculating the characteristic value lambda of R1,λ2,...λmAnd a feature vector u corresponding to each feature value1,u2,...umThen, the variance contribution rate eta is calculatediAnd the cumulative variance contribution η (p).
Figure BDA0003371957350000093
Figure BDA0003371957350000094
And selecting p which can enable the cumulative variance contribution rate to be larger than 75% as the number of main components, namely the dimension of the feature vector after dimension reduction. Finally, dimension reduction from the m-dimensional voiceprint feature vector to the p-dimensional feature vector is achieved.
S102: and inputting the characteristic vector serving as an input parameter into a preset voiceprint recognition model to obtain a working condition label corresponding to the voiceprint signal to be detected.
And the preset voiceprint recognition model is subjected to model training based on transfer learning through a voiceprint feature analysis network, and voiceprint recognition models of the substation equipment under different working conditions are established.
In the embodiment of the application, after the voiceprint features of the power transformation equipment are extracted, the working condition of the power transformation equipment needs to be modeled so as to compare and judge the voiceprint signals to be detected. The method and the device for identifying the voiceprint signals are based on DNN, and are improved aiming at the voiceprint characteristics of the power transformation equipment, so that voiceprint signal identification is achieved.
The structure of the DNN-based transformer voiceprint feature analysis network is shown in fig. 6. In the network structure, the input is the result of splicing the extracted frame characteristic vectors, and the output is One-hot vectors corresponding to different working conditions. ReLU is used as an activation function in the network, SoftMax function is used for classification, and cross entropy function is used as an objective function.
The training based on the transfer learning is based on the common characteristic of the acoustic characteristics, and the voiceprint signal characteristics of the power transformation equipment of different models are trained and learned by using the thought of parameter transfer. Specifically, a pre-trained model trained using a large amount of generic data is trained in a manner that fits a particular monitored target. After training the neural network with generic voiceprint data, a basic voiceprint network framework is obtained as shown in fig. 7. The network has 4 hidden layers in total, the parameters of layers 1-2 of the trained model are reserved, and the parameters of layers 3-4 are continuously trained on the basis; after the second training, all parameters are fixed, the last classification layer is removed, and the last output layer of the hidden layer is used as a template of a monitoring target for comparison.
Further, the template selection and measurement index of the voiceprint recognition is used, after the training data is used for training the neural network, the last output classification layer with the label is removed through the trained neural network with fixed parameters, and the output of the last hidden layer is used as the voiceprint template corresponding to the working condition. And for different working conditions, randomly taking out 5-15 data in the training data set, and obtaining the target template through a neural network. Because the template selection has certain randomness, in order to ensure that the target template can accurately reflect the voiceprint conditions under the working conditions, the obtained target template is clustered by using a k-means algorithm. Specifically, clustering with k being 2 is performed on the 5-15 corresponding target template outputs; and after finishing clustering, selecting the cluster where most samples are positioned as a sample template representing the working condition.
S103: and performing similarity comparison on the working condition label and the target template output by utilizing a cosine similarity measurement algorithm formula so as to realize intelligent voiceprint fault detection of the substation equipment.
In the embodiment of the application, the voiceprint signal to be detected is sent to a trained DNN model after extracting the characteristics of the voiceprint signal to be detected, and then the output o of the last hidden layer is obtained. And comparing the similarity between o and the target template output p to realize voiceprint fault detection of the power transformation equipment, wherein specifically, the cosine similarity measurement algorithm formula is as follows:
Figure BDA0003371957350000111
the embodiment of the application provides a voiceprint fault intelligent detection method for substation equipment, which comprises the steps of after obtaining a voiceprint signal to be detected, carrying out voiceprint feature extraction on the voiceprint signal to be detected to obtain a feature vector corresponding to the voiceprint signal to be detected; inputting the characteristic vector serving as an input parameter into a preset voiceprint recognition model to obtain a working condition label corresponding to the voiceprint signal to be detected, carrying out model training based on transfer learning by the preset voiceprint recognition model through a voiceprint feature analysis network, and establishing voiceprint recognition models of the substation equipment under different working conditions; and performing similarity comparison on the working condition label and the target template output by utilizing a cosine similarity measurement algorithm formula so as to realize intelligent voiceprint fault detection of the substation equipment. Through this application embodiment can effectively promote the voiceprint fault detection efficiency of substation equipment, reduce the manual work and patrol and examine the cost to promote the intelligent level that substation equipment patrolled and examined comprehensively.
Referring to fig. 8, based on the voiceprint fault intelligent detection method for the substation equipment disclosed in the foregoing embodiment, this embodiment correspondingly discloses a voiceprint fault intelligent detection device for the substation equipment, and the device includes:
the first processing unit 801 is configured to acquire a voiceprint signal to be detected, and perform voiceprint feature extraction on the voiceprint signal to be detected to obtain a feature vector corresponding to the voiceprint signal to be detected;
the second processing unit 802 is configured to input the feature vectors as input parameters to a preset voiceprint recognition model to obtain a working condition label corresponding to the voiceprint signal to be detected, where the preset voiceprint recognition model performs model training based on transfer learning through a voiceprint feature analysis network, and establishes voiceprint recognition models of the substation equipment under different working conditions;
and the third processing unit 803 is configured to perform similarity comparison on the operating condition tag and the target template output by using a cosine similarity measurement algorithm formula, so as to implement intelligent voiceprint fault detection of the substation equipment.
Further, the first processing unit 801 is configured to:
acquiring a voiceprint signal to be detected, and preprocessing the voiceprint signal to be detected to obtain a plurality of framing signals;
respectively extracting short-time domain features and short-time frequency domain features from each frame signal, and generating an MFCC feature vector group, wherein the short-time domain features comprise short-time energy, a short-time average zero crossing rate and a short-time energy entropy, and the short-time frequency domain features are Mel frequency cepstrum coefficients;
and performing characteristic connection on the MFCC characteristic vectors in the MFCC characteristic vector group to obtain a characteristic vector corresponding to the voiceprint signal to be detected.
Further, the first processing unit 801 is further configured to:
and introducing a weighting optimization method based on an F ratio into the MFCC feature vector group, and performing dimension reduction processing on the MFCC feature vectors in the MFCC feature vector group based on multi-modal features of PCA.
Further, the first processing unit 801 is further configured to:
dividing the voiceprint signal to be detected collected in a preset time period into a plurality of short-time signals;
and after windowing each short-time signal by adopting a Hamming window, carrying out Fourier transform on each short-time signal to obtain a plurality of framing signals.
The voiceprint fault intelligent detection device for the substation equipment comprises a processor and a memory, wherein the first processing unit, the second processing unit, the third processing unit and the like are stored in the memory as program units, and the processor executes the program units stored in the memory to realize corresponding functions.
The processor comprises a kernel, and the kernel calls the corresponding program unit from the memory. The kernel can set up one or more, effectively promotes transformer substation equipment's vocal print fault detection efficiency through adjusting kernel parameter, reduces the manual work and patrols and examines the cost to promote the intelligent level that transformer substation equipment patrolled and examined comprehensively.
The embodiment of the application provides a storage medium, wherein a program is stored on the storage medium, and when the program is executed by a processor, the intelligent voiceprint fault detection method for the substation equipment is realized.
The embodiment of the application provides a processor, wherein the processor is used for running a program, and the voiceprint fault intelligent detection method for substation equipment is executed when the program runs.
An electronic device is provided in the embodiments of the present application, as shown in fig. 9, the electronic device 90 includes at least one processor 901, at least one memory 902 connected to the processor, and a bus 903; the processor 901 and the memory 902 complete communication with each other through the bus 903; the processor 901 is configured to call the program instructions in the memory 902 to execute the voiceprint fault intelligent detection method for the substation equipment.
The electronic device herein may be a server, a PC, a PAD, a mobile phone, etc.
The present application further provides a computer program product adapted to perform a program for initializing the following method steps when executed on a data processing device:
acquiring a voiceprint signal to be detected, and extracting voiceprint characteristics of the voiceprint signal to be detected to obtain a characteristic vector corresponding to the voiceprint signal to be detected;
inputting the characteristic vector serving as an input parameter into a preset voiceprint recognition model to obtain a working condition label corresponding to the voiceprint signal to be detected, carrying out model training based on transfer learning by the preset voiceprint recognition model through a voiceprint feature analysis network, and establishing voiceprint recognition models of the substation equipment under different working conditions;
and performing similarity comparison on the working condition label and the target template output by utilizing a cosine similarity measurement algorithm formula so as to realize intelligent voiceprint fault detection of the substation equipment.
Further, the acquiring a voiceprint signal to be detected, and performing voiceprint feature extraction on the voiceprint signal to be detected to obtain a feature vector corresponding to the voiceprint signal to be detected includes:
acquiring a voiceprint signal to be detected, and preprocessing the voiceprint signal to be detected to obtain a plurality of framing signals;
respectively extracting short-time domain features and short-time frequency domain features from each frame signal, and generating an MFCC feature vector group, wherein the short-time domain features comprise short-time energy, a short-time average zero crossing rate and a short-time energy entropy, and the short-time frequency domain features are Mel frequency cepstrum coefficients;
and performing characteristic connection on the MFCC characteristic vectors in the MFCC characteristic vector group to obtain a characteristic vector corresponding to the voiceprint signal to be detected.
Further, before the feature concatenating the MFCC feature vectors in the MFCC feature vector group, the method further includes:
and introducing a weighting optimization method based on an F ratio into the MFCC feature vector group, and performing dimension reduction processing on the MFCC feature vectors in the MFCC feature vector group based on multi-modal features of PCA.
Further, the preprocessing the voiceprint signal to be detected to obtain a plurality of framing signals includes:
dividing the voiceprint signal to be detected collected in a preset time period into a plurality of short-time signals;
and after windowing each short-time signal by adopting a Hamming window, carrying out Fourier transform on each short-time signal to obtain a plurality of framing signals.
The present application is described in terms of flowcharts and/or block diagrams of methods, apparatus (systems), computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In a typical configuration, a device includes one or more processors (CPUs), memory, and a bus. The device may also include input/output interfaces, network interfaces, and the like.
The memory may include volatile memory in a computer readable medium, Random Access Memory (RAM) and/or nonvolatile memory such as Read Only Memory (ROM) or flash memory (flash RAM), and the memory includes at least one memory chip. The memory is an example of a computer-readable medium.
Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in the process, method, article, or apparatus that comprises the element.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The above are merely examples of the present application and are not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims (10)

1. A voiceprint fault intelligent detection method for substation equipment is characterized by comprising the following steps:
acquiring a voiceprint signal to be detected, and extracting voiceprint characteristics of the voiceprint signal to be detected to obtain a characteristic vector corresponding to the voiceprint signal to be detected;
inputting the characteristic vector serving as an input parameter into a preset voiceprint recognition model to obtain a working condition label corresponding to the voiceprint signal to be detected, carrying out model training based on transfer learning by the preset voiceprint recognition model through a voiceprint feature analysis network, and establishing voiceprint recognition models of the substation equipment under different working conditions;
and performing similarity comparison on the working condition label and the target template output by utilizing a cosine similarity measurement algorithm formula so as to realize intelligent voiceprint fault detection of the substation equipment.
2. The method according to claim 1, wherein the obtaining of the voiceprint signal to be detected and the voiceprint feature extraction of the voiceprint signal to be detected to obtain the feature vector corresponding to the voiceprint signal to be detected comprises:
acquiring a voiceprint signal to be detected, and preprocessing the voiceprint signal to be detected to obtain a plurality of framing signals;
respectively extracting short-time domain features and short-time frequency domain features from each frame signal, and generating an MFCC feature vector group, wherein the short-time domain features comprise short-time energy, a short-time average zero crossing rate and a short-time energy entropy, and the short-time frequency domain features are Mel frequency cepstrum coefficients;
and performing characteristic connection on the MFCC characteristic vectors in the MFCC characteristic vector group to obtain a characteristic vector corresponding to the voiceprint signal to be detected.
3. The method of claim 2, further comprising, prior to said feature concatenating MFCC feature vectors in the set of MFCC feature vectors:
and introducing a weighting optimization method based on an F ratio into the MFCC feature vector group, and performing dimension reduction processing on the MFCC feature vectors in the MFCC feature vector group based on multi-modal features of PCA.
4. The method according to claim 2, wherein the preprocessing the voiceprint signal to be detected to obtain a plurality of framing signals comprises:
dividing the voiceprint signal to be detected collected in a preset time period into a plurality of short-time signals;
and after windowing each short-time signal by adopting a Hamming window, carrying out Fourier transform on each short-time signal to obtain a plurality of framing signals.
5. The utility model provides a voiceprint fault intelligent detection device towards substation equipment which characterized in that includes:
the first processing unit is used for acquiring a voiceprint signal to be detected and extracting voiceprint characteristics of the voiceprint signal to be detected to obtain a characteristic vector corresponding to the voiceprint signal to be detected;
the second processing unit is used for inputting the characteristic vectors serving as input parameters into a preset voiceprint recognition model to obtain a working condition label corresponding to the voiceprint signal to be detected, and the preset voiceprint recognition model is used for carrying out model training based on transfer learning through a voiceprint feature analysis network to establish voiceprint recognition models of the substation equipment under different working conditions;
and the third processing unit is used for comparing the similarity of the working condition label and the target template output by utilizing a cosine similarity measurement algorithm formula so as to realize the intelligent voiceprint fault detection of the substation equipment.
6. The apparatus of claim 5, wherein the first processing unit is configured to:
acquiring a voiceprint signal to be detected, and preprocessing the voiceprint signal to be detected to obtain a plurality of framing signals;
respectively extracting short-time domain features and short-time frequency domain features from each frame signal, and generating an MFCC feature vector group, wherein the short-time domain features comprise short-time energy, a short-time average zero crossing rate and a short-time energy entropy, and the short-time frequency domain features are Mel frequency cepstrum coefficients;
and performing characteristic connection on the MFCC characteristic vectors in the MFCC characteristic vector group to obtain a characteristic vector corresponding to the voiceprint signal to be detected.
7. The apparatus of claim 6, wherein the first processing unit is further configured to:
and introducing a weighting optimization method based on an F ratio into the MFCC feature vector group, and performing dimension reduction processing on the MFCC feature vectors in the MFCC feature vector group based on multi-modal features of PCA.
8. The apparatus of claim 6, wherein the first processing unit is further configured to:
dividing the voiceprint signal to be detected collected in a preset time period into a plurality of short-time signals;
and after windowing each short-time signal by adopting a Hamming window, carrying out Fourier transform on each short-time signal to obtain a plurality of framing signals.
9. A storage medium, characterized in that the storage medium comprises a stored program, wherein when the program runs, a device in which the storage medium is located is controlled to execute the voiceprint fault intelligent detection method for substation equipment according to any one of claims 1 to 4.
10. An electronic device comprising at least one processor, and at least one memory, bus connected to the processor; the processor and the memory complete mutual communication through the bus; the processor is used for calling the program instructions in the memory to execute the voiceprint fault intelligent detection method for the substation equipment according to any one of claims 1 to 4.
CN202111404981.2A 2021-11-24 2021-11-24 Voiceprint fault intelligent detection method and device for substation equipment Pending CN114121025A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111404981.2A CN114121025A (en) 2021-11-24 2021-11-24 Voiceprint fault intelligent detection method and device for substation equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111404981.2A CN114121025A (en) 2021-11-24 2021-11-24 Voiceprint fault intelligent detection method and device for substation equipment

Publications (1)

Publication Number Publication Date
CN114121025A true CN114121025A (en) 2022-03-01

Family

ID=80372147

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111404981.2A Pending CN114121025A (en) 2021-11-24 2021-11-24 Voiceprint fault intelligent detection method and device for substation equipment

Country Status (1)

Country Link
CN (1) CN114121025A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111261189A (en) * 2020-04-02 2020-06-09 中国科学院上海微系统与信息技术研究所 Vehicle sound signal feature extraction method

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111261189A (en) * 2020-04-02 2020-06-09 中国科学院上海微系统与信息技术研究所 Vehicle sound signal feature extraction method

Similar Documents

Publication Publication Date Title
CN112201260B (en) Transformer running state online detection method based on voiceprint recognition
CN109034046B (en) Method for automatically identifying foreign matters in electric energy meter based on acoustic detection
CN109949823B (en) DWPT-MFCC and GMM-based in-vehicle abnormal sound identification method
CN112885372B (en) Intelligent diagnosis method, system, terminal and medium for power equipment fault sound
CN111048114A (en) Equipment and method for detecting abnormal sound of equipment
CN111325095A (en) Intelligent equipment health state detection method and system based on sound wave signals
CN110120230B (en) Acoustic event detection method and device
CN112420055A (en) Substation state identification method and device based on voiceprint characteristics
CN112329914B (en) Fault diagnosis method and device for buried transformer substation and electronic equipment
CN111724770A (en) Audio keyword identification method for generating confrontation network based on deep convolution
CN114023354A (en) Guidance type acoustic event detection model training method based on focusing loss function
Kim et al. Hierarchical approach for abnormal acoustic event classification in an elevator
CN112435686A (en) Power equipment fault voice recognition method based on data enhancement
CN112599134A (en) Transformer sound event detection method based on voiceprint recognition
CN116778964A (en) Power transformation equipment fault monitoring system and method based on voiceprint recognition
CN114121025A (en) Voiceprint fault intelligent detection method and device for substation equipment
CN114974229A (en) Method and system for extracting abnormal behaviors based on audio data of power field operation
CN107894837A (en) Dynamic sentiment analysis model sample processing method and processing device
CN116773952A (en) Transformer voiceprint signal fault diagnosis method and system
CN116884435A (en) Voice event detection method and device based on audio prompt learning
CN114818832A (en) Multi-scale feature fusion transformer voiceprint classification method
He et al. Deep learning approach for audio signal classification and its application in fiber optic sensor security system
CN112232329A (en) Multi-core SVM training and alarming method, device and system for intrusion signal recognition
CN116072146A (en) Pumped storage station detection method and system based on voiceprint recognition
CN117975999A (en) Valve cooling equipment fault detection method and device, computer equipment and storage medium

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