CN117217730A - Power equipment fault identification method, device, equipment, medium and product - Google Patents
Power equipment fault identification method, device, equipment, medium and product Download PDFInfo
- Publication number
- CN117217730A CN117217730A CN202311088522.7A CN202311088522A CN117217730A CN 117217730 A CN117217730 A CN 117217730A CN 202311088522 A CN202311088522 A CN 202311088522A CN 117217730 A CN117217730 A CN 117217730A
- Authority
- CN
- China
- Prior art keywords
- voiceprint
- data
- power equipment
- fault
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 63
- 238000012549 training Methods 0.000 claims abstract description 46
- 230000009466 transformation Effects 0.000 claims abstract description 31
- 238000004458 analytical method Methods 0.000 claims description 69
- 238000004590 computer program Methods 0.000 claims description 22
- 238000001228 spectrum Methods 0.000 claims description 12
- 238000007781 pre-processing Methods 0.000 claims description 11
- 238000000605 extraction Methods 0.000 claims description 6
- 238000009432 framing Methods 0.000 claims description 5
- 238000009499 grossing Methods 0.000 claims description 5
- 238000002372 labelling Methods 0.000 claims description 3
- 238000007689 inspection Methods 0.000 abstract description 6
- 238000012217 deletion Methods 0.000 abstract 1
- 230000037430 deletion Effects 0.000 abstract 1
- 238000007726 management method Methods 0.000 description 20
- 239000013598 vector Substances 0.000 description 18
- 238000004422 calculation algorithm Methods 0.000 description 14
- 238000012423 maintenance Methods 0.000 description 14
- 230000006870 function Effects 0.000 description 12
- 238000013528 artificial neural network Methods 0.000 description 11
- 238000012545 processing Methods 0.000 description 10
- 238000003745 diagnosis Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 9
- 238000012544 monitoring process Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 5
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 238000012360 testing method Methods 0.000 description 4
- 238000010276 construction Methods 0.000 description 3
- 238000013523 data management Methods 0.000 description 3
- 238000013500 data storage Methods 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 239000007787 solid Substances 0.000 description 3
- 238000011426 transformation method Methods 0.000 description 3
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000004913 activation Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000015556 catabolic process Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000006731 degradation reaction Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000004807 localization Effects 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000002035 prolonged effect Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000003213 activating effect Effects 0.000 description 1
- 238000013475 authorization Methods 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005684 electric field Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 229910021389 graphene Inorganic materials 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application relates to a power equipment fault identification method, a device, equipment, a medium and a product. The method comprises the following steps: receiving initial voiceprint data of the power equipment sent by a voiceprint acquisition terminal; performing feature transformation on the initial voiceprint data to obtain time-frequency domain feature data, and obtaining fault labels corresponding to the time-frequency domain feature data based on the working states of the power equipment in the time period corresponding to the initial voiceprint data; and performing parameter tuning on the platform fault identification model by using a training sample to obtain tuned model parameters, and transmitting the model parameters to the voiceprint acquisition terminal, so that the voiceprint acquisition terminal determines a terminal fault identification model built in the voiceprint acquisition terminal according to the model parameters after receiving the model parameters, and performs fault identification on voiceprint information to be identified of the power equipment, which is locally and in real time acquired by the terminal, through the terminal fault identification model. By adopting the method, the inspection efficiency of the power equipment can be improved, and the data falsification or deletion can be avoided.
Description
Technical Field
The present application relates to the field of fault identification technologies, and in particular, to a method, an apparatus, a device, a medium, and a product for identifying a fault of an electrical device.
Background
The middle-low voltage distribution electric equipment is a weak link of power grid operation control and operation maintenance management for a long time, and has the advantages of wide points and multiple involved links, taking a distribution transformer as an example, being an important electric equipment, being closely related with end users, being numerous in number and wide in distribution range, and being subjected to long-time power failure once a fault or planned outage occurs. According to statistics, the distribution transformer accidents account for 90% of the whole distribution network accidents.
Along with the rising of the edge data center, related power distribution equipment is various, real-time monitoring of the running state of the power distribution equipment has a critical effect on the power supply reliability, and equipment faults can be detected and repaired in time in a targeted manner, so that the operation and maintenance efficiency is improved. However, the current monitoring and inspection means are low in efficiency, have the problems of blind control, blind adjustment and the like, and most of power equipment is inspected manually by inspection personnel, so that the efficiency is low, and the risk of data falsification or missing easily occurs.
Disclosure of Invention
In view of the foregoing, it is desirable to provide a method, apparatus, device, medium, and product for identifying a power equipment failure, which can improve inspection efficiency of the power equipment and avoid data falsification or loss.
In a first aspect, the present application provides a method for identifying a power device fault. The method comprises the following steps:
receiving initial voiceprint data of the power equipment sent by a voiceprint acquisition terminal;
performing feature transformation on the initial voiceprint data to obtain time-frequency domain feature data, and obtaining fault labels corresponding to the time-frequency domain feature data based on the working state of the power equipment in a time period corresponding to the initial voiceprint data, wherein the fault labels comprise normal power equipment and power equipment faults;
and carrying out parameter tuning on a preset platform fault identification model by using a training sample to obtain tuned model parameters, and transmitting the tuned model parameters to a voiceprint acquisition terminal so that the voiceprint acquisition terminal can determine a terminal fault identification model built in the voiceprint acquisition terminal according to the tuned model parameters after receiving the tuned model parameters, and carrying out fault identification on voiceprint information to be identified of the power equipment, which is locally and in real time acquired by the terminal, through the terminal fault identification model, wherein the training sample comprises time-frequency domain feature data and corresponding records of fault labels.
In one embodiment, performing feature transformation on the initial voiceprint data to obtain time-frequency domain feature data, including:
Normalizing the initial voiceprint data to obtain normalized data;
performing wavelet transformation on the normalized data based on preset super parameters to obtain time frequency spectrums on different wavelet sub-bands;
and extracting cepstrum coefficients of the time spectrum on each wavelet sub-band according to the time spectrum on each wavelet sub-band to obtain time-frequency domain characteristic data.
In one embodiment, after receiving the initial voiceprint data of the power device sent by the voiceprint acquisition terminal, the method further includes:
preprocessing the initial voiceprint data to update the initial voiceprint data;
wherein the preprocessing includes at least one of pre-emphasis, framing, blank rejection, and active frame smoothing.
In one embodiment, initial voiceprint data of different power devices in different time periods, which are sent by different voiceprint acquisition terminals, are received;
acquiring a plurality of corresponding records between time-frequency domain characteristic data corresponding to different initial voiceprint data and a fault label, and storing the corresponding records;
responding to a demand instruction of a user, selecting at least one corresponding record from a plurality of corresponding records to construct a training sample set, and performing parameter tuning on a preset platform fault identification model through the training sample set.
In a second aspect, the present application further provides a method for identifying a fault of an electrical device, which is applied to a voiceprint acquisition terminal, where the method includes:
acquiring initial voiceprint data of the power equipment, and sending the initial voiceprint data to a voiceprint analysis platform;
receiving model parameters sent by a voiceprint analysis platform, and determining a terminal fault identification model built in a voiceprint acquisition terminal according to the model parameters, wherein the model parameters are parameters obtained after the voiceprint analysis platform performs characteristic transformation on initial voiceprint data to obtain time-frequency domain characteristic data, and obtains fault labels corresponding to the time-frequency domain characteristic data based on the working state of power equipment in a time period corresponding to the initial voiceprint data, and parameter tuning is performed on the preset platform fault identification model by utilizing the time-frequency domain characteristic data and the corresponding records of the fault labels to obtain optimized parameters;
and carrying out fault recognition on the voiceprint information to be recognized of the power equipment, which is locally and in real time acquired by the terminal, through the terminal fault recognition model.
In a third aspect, the application further provides a device for identifying faults of the power equipment. The device comprises:
the receiving data module is used for receiving initial voiceprint data of the power equipment sent by the voiceprint acquisition terminal;
The extraction labeling module is used for carrying out feature transformation on the initial voiceprint data to obtain time-frequency domain feature data, and obtaining fault labels corresponding to the time-frequency domain feature data based on the working state of the power equipment in a time period corresponding to the initial voiceprint data, wherein the fault labels comprise normal power equipment and power equipment faults;
the tuning parameter module is used for performing parameter tuning on a preset platform fault identification model by using a training sample to obtain tuned model parameters, and transmitting the tuned model parameters to the voiceprint acquisition terminal, so that the voiceprint acquisition terminal can determine a terminal fault identification model built in the voiceprint acquisition terminal according to the tuned model parameters after receiving the tuned model parameters, and perform fault identification on voiceprint information to be identified of the power equipment, which is locally and in real time acquired by the terminal, through the terminal fault identification model, wherein the training sample comprises time-frequency domain feature data and corresponding records of fault labels.
In a fourth aspect, the application further provides a power equipment fault identification device. The device comprises:
the acquisition data module is used for acquiring initial voiceprint data of the power equipment and sending the initial voiceprint data to the voiceprint analysis platform;
The system comprises a receiving parameter module, a voice print analysis platform and a fault label processing module, wherein the receiving parameter module is used for receiving model parameters sent by the voice print analysis platform and determining a terminal fault identification model built in the voice print acquisition terminal according to the model parameters, the model parameters are obtained by performing characteristic transformation on initial voice print data by the voice print analysis platform, obtaining time-frequency domain characteristic data, obtaining fault labels corresponding to the time-frequency domain characteristic data based on the working state of power equipment in a time period corresponding to the initial voice print data, and performing parameter tuning on the preset platform fault identification model by utilizing the time-frequency domain characteristic data and the corresponding records of the fault labels to obtain tuned parameters;
and the fault recognition module is used for carrying out fault recognition on the voice print information to be recognized of the power equipment, which is locally and in real time acquired by the terminal, through the terminal fault recognition model.
In a fifth aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor implementing the power device fault identification method provided in the first or second aspect of the application when the processor executes the computer program.
In a sixth aspect, the present application also provides a computer readable storage medium. The computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements the power equipment fault identification method provided in the first aspect or the second aspect of the present application.
In a seventh aspect, the present application also provides a computer program product. The computer program product comprises a computer program which, when executed by a processor, implements the power equipment fault identification method provided in the first or second aspect of the present application.
According to the power equipment fault identification method, the device, the equipment, the medium and the product, initial voiceprint data of the power equipment transmitted by the voiceprint acquisition terminal are received, characteristic transformation is carried out on the initial voiceprint data, time-frequency domain characteristic data are obtained, fault labels corresponding to the time-frequency domain characteristic data are obtained based on the working states of the power equipment in a time period corresponding to the initial voiceprint data, a training sample is utilized to carry out parameter tuning on a preset platform fault identification model, model parameters after tuning are obtained, the model parameters after tuning are issued to the voiceprint acquisition terminal, so that after the voiceprint acquisition terminal receives the model parameters after tuning, a terminal fault identification model built in the voiceprint acquisition terminal is determined according to the model parameters after tuning, and fault identification is carried out on the power equipment to be identified locally and in real time through the terminal fault identification model. Compared with signals such as voltage, current, oil temperature and the like which are widely adopted at present, the voice signal has the characteristics of low acquisition cost, small influence on equipment, convenient construction and installation and the like, and the relationship between voice print characteristics and abnormal states of the equipment is close, the voice print acquisition terminal and the voice print analysis cloud platform are deployed, voice print audio information of power equipment operation is acquired in real time by using the voice print acquisition terminal, time-frequency domain characteristics of the voice print information are extracted and marked by using the voice print analysis cloud platform through a characteristic transformation method, training samples are obtained through time-frequency domain characteristic data and marked labels, a platform fault identification model is trained, adjusted model parameters are obtained, the model parameters are issued to the voice print acquisition terminal, real-time localization edge fault identification is realized by the voice print acquisition terminal, and a solid foundation is provided for adapting to more various Internet of things equipment and complex Internet of things environments. The work of the patrol workers is greatly reduced or completely replaced, repeated and boring invalid operation of the patrol workers is reduced, and the dependence on the quality and the enthusiasm of the patrol workers is greatly reduced; the equipment state monitoring data are real-time, reliable and safe, and the safety risks and the data fake or missing risks brought by personnel inspection are reduced.
Furthermore, the embodiment of the application can also monitor the working state of the power equipment in real time, discover the degradation trend of the power equipment at the first time, and provide plentiful time for making maintenance decisions; the state of the power equipment is controlled and predictably maintained, the maintenance interval is prolonged, and technical support is provided for reasonably reducing maintenance cost; avoiding unexpected shutdown until the unplanned shutdown is stopped, and avoiding yield loss caused by unexpected shutdown; and equipment faults and fault reasons are found in time, equipment overhaul guidance is provided for an overhaul party, overhaul time is shortened, and overhaul cost is reduced. The fault part is prejudged in advance, the warehouse is checked in advance, and the problem that standard accessories cannot be found during maintenance, so that people wait for spare parts and waste maintenance time is avoided; or other damage to the equipment in the event of replacement with other emergency spare parts; and evaluating equipment or giving physical reports periodically, optimizing inventory and purchasing plans according to the reports, and optimizing the quantity and structure of spare parts.
Drawings
FIG. 1 is an application environment diagram of a power device fault identification method in one embodiment;
FIG. 2 is a flow chart of a method for identifying a power device failure in one embodiment;
FIG. 3 is a schematic diagram of an architecture of a voiceprint acquisition terminal in one embodiment;
FIG. 4 is a diagram of a model structure of a deep neural network in one embodiment;
FIG. 5 is a schematic diagram of an architecture of a voiceprint analysis cloud platform in one embodiment;
FIG. 6 is a schematic diagram of a voiceprint analysis cloud platform and voiceprint acquisition terminal in another embodiment;
FIG. 7 is a flow chart of a method for identifying a power device failure in another embodiment;
FIG. 8 is a block diagram of a power equipment failure recognition device in one embodiment;
FIG. 9 is a block diagram of a power equipment failure recognition device in another embodiment;
fig. 10 is an internal structural view of a computer device in one embodiment.
Detailed Description
The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.
For a long time, the acoustic signals generated by electrical equipment have been generally ignored as noise. In practical situations, in the running process of the electric power equipment, mechanical vibration can occur to equipment elements due to the action of magnetostriction force and electric field force, and acoustic signals of the electric power equipment are generated through the propagation of an air medium, so that a large amount of state information of the electric power equipment is contained. When the power equipment fails, the internal structure changes, so that the acoustic signal changes, and the power equipment can be subjected to fault diagnosis through analysis of the acoustic signal.
The power equipment fault identification method provided by the embodiment of the application can be applied to an application environment shown in figure 1. The voiceprint acquisition terminal 102 communicates with the voiceprint analysis cloud platform 104 through a network. The data storage system may store data that the voiceprint analysis cloud platform 104 needs to process. The data storage system may be integrated on the voiceprint analysis cloud platform 104 or may be located on other network servers. The voiceprint acquisition terminal 102 can be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, internet of things devices, portable wearable devices, and the internet of things devices can be smart speakers, smart televisions, smart air conditioners, smart vehicle devices, and the like. The portable wearable device may be a smart watch, smart bracelet, headset, or the like. The voiceprint analysis cloud platform 104 can be implemented with a stand-alone server or a server cluster of multiple servers.
In one embodiment, as shown in fig. 2, a method for identifying a fault of an electrical device is provided, and the method is applied to the voiceprint analysis cloud platform in fig. 1 for illustration, and includes the following steps:
step 202, initial voiceprint data of power equipment sent by a voiceprint acquisition terminal is received.
The voiceprint acquisition terminal is a device which is arranged around the power equipment and used for acquiring voiceprint information of the power equipment in real time.
The voiceprint acquisition terminals can acquire initial voiceprint data of the power equipment in real time, wherein a plurality of voiceprint acquisition terminals can be provided, each voiceprint acquisition terminal can be used for acquiring voiceprint data of at least one type of power equipment, and each voiceprint acquisition terminal can also acquire voiceprint data of different time periods.
For example, the voiceprint acquisition terminal may include a first voiceprint acquisition terminal and a second voiceprint acquisition terminal, the first voiceprint acquisition terminal may acquire different initial voiceprint data corresponding to the first air conditioner in different time periods, different initial voiceprint data corresponding to the first transformer in different time periods, and the second voiceprint acquisition terminal may acquire different initial voiceprint data corresponding to the second air conditioner in different time periods, and different initial voiceprint data corresponding to the second transformer in different time periods.
After the voiceprint acquisition terminal acquires the initial voiceprint data, the initial voiceprint data is uploaded to the voiceprint analysis cloud platform, namely the voiceprint analysis cloud platform receives the initial voiceprint data sent by the voiceprint acquisition terminal.
And 204, performing feature transformation on the initial voiceprint data to obtain time-frequency domain feature data, and obtaining a fault tag corresponding to the time-frequency domain feature data based on the working state of the power equipment in the time period corresponding to the initial voiceprint data.
The time-frequency domain feature data refers to feature data extracted by analyzing initial voiceprint data on a time axis and a frequency axis.
After receiving the initial voiceprint data, the embodiment of the application can acquire the time-frequency domain feature data corresponding to the initial voiceprint data in a feature transformation mode, for example, the feature transformation can adopt at least one of principal component analysis, fourier transformation, wavelet packet transformation, local feature extraction and the like. After the time-frequency domain feature data corresponding to the initial voiceprint data is determined, the time-frequency domain feature data can be marked, and a fault label corresponding to the time-frequency domain feature data is obtained based on the working state of the power equipment in a time period corresponding to the initial voiceprint data, wherein the fault label comprises the normal power equipment and the fault of the power equipment.
In one implementation manner, the voiceprint analysis cloud platform can receive initial voiceprint data, which are sent by different voiceprint acquisition terminals and correspond to different power equipment in different time periods, and after feature transformation is performed on the different initial voiceprint data to obtain time-frequency domain feature data and label the time-frequency domain feature data, corresponding records of a plurality of pieces of time-frequency domain feature data and fault labels can be stored on the cloud platform, and a user can conduct management operations such as export, input, modification and the like on the time-frequency domain feature data.
And 206, performing parameter tuning on a preset platform fault recognition model by using a training sample to obtain tuned model parameters, and transmitting the tuned model parameters to the voiceprint acquisition terminal so that the voiceprint acquisition terminal can determine a terminal fault recognition model built in the voiceprint acquisition terminal according to the tuned model parameters after receiving the tuned model parameters, and performing fault recognition on the voiceprint information to be recognized of the power equipment, which is locally and in real time acquired by the terminal, through the terminal fault recognition model.
The training samples comprise corresponding records of time-frequency domain characteristic data and fault labels.
According to the embodiment of the application, the corresponding records of the time-frequency domain characteristic data and the fault label are used as training samples, the parameter tuning is performed on the preset platform fault recognition model, specifically, the time-frequency domain characteristic data can be input into the preset platform fault recognition model, the normal or fault working state of the power equipment is output, the loss value is calculated through the output result and the fault label, the model parameter is tuned, and the tuned model parameter is obtained.
In one implementation manner, the voiceprint analysis cloud platform can store time-frequency domain feature data and corresponding fault labels corresponding to initial voiceprint data of different power equipment transmitted by different voiceprint acquisition terminals in different time periods, and can select different time-frequency domain feature data to construct a training sample set and a test sample set according to different base structures of the platform fault identification model, different types of the power equipment to be identified, different data timeliness and the like, and after the base parameters of the platform fault identification model are initialized, the training sample set is used for training the platform fault identification model, the test sample set is used for testing accuracy, and model parameters of the platform fault identification model which is completed through training are obtained. The model parameters are stored on the cloud platform, and management operations such as export, input, modification and the like can be performed on the model parameters.
Further, platform fault recognition models of different basic structures or different training sample sets can be adopted for training to obtain model parameters which are adjusted and optimized under different conditions, and specific model parameters are issued to the voiceprint acquisition terminal according to user requirements.
The voiceprint acquisition terminal of the embodiment of the application utilizes the model parameters issued by the voiceprint analysis cloud platform after tuning to construct a terminal fault identification model of the terminal side. The voiceprint acquisition terminal can acquire voiceprint information to be identified of the power equipment in real time, time-frequency domain feature data corresponding to the voiceprint information to be identified are obtained through corresponding feature transformation processing, then the time-frequency domain feature data corresponding to the voiceprint information to be identified are input into the terminal fault identification model for fault identification, and potential fault risks of the equipment are identified, so that the working state of the power equipment is monitored in real time.
In the above power equipment fault identification method, initial voiceprint data of the power equipment transmitted by the voiceprint acquisition terminal is received, characteristic transformation is performed on the initial voiceprint data, time-frequency domain characteristic data are obtained, fault labels corresponding to the time-frequency domain characteristic data are obtained based on the working states of the power equipment in a time period corresponding to the initial voiceprint data, a training sample is utilized to perform parameter tuning on a preset platform fault identification model, model parameters after tuning are obtained, the model parameters after tuning are issued to the voiceprint acquisition terminal, after the voiceprint acquisition terminal receives the model parameters after tuning, a terminal fault identification model built in the voiceprint acquisition terminal is determined according to the model parameters after tuning, and fault identification is performed on power equipment to be identified locally and in real time through the terminal fault identification model. Compared with signals such as voltage, current and oil temperature which are widely adopted at present, the voice signal has the characteristics of low acquisition cost, small influence on equipment, convenient construction and installation and the like, and the relationship between voice print characteristics and abnormal states of the equipment is close, the voice print acquisition terminal and the voice print analysis cloud platform are deployed, voice print audio information of power equipment operation is acquired in real time by using the voice print acquisition terminal, time-frequency domain characteristics of the voice print information are extracted and marked by using the voice print analysis cloud platform through a characteristic transformation method, training samples are obtained through time-frequency domain characteristic data and marked labels, a platform fault identification model is trained, adjusted model parameters are obtained, the model parameters are issued to the voice print acquisition terminal, real-time localization edge fault identification is realized by the voice print acquisition terminal, and a solid foundation is provided for adapting to more various Internet of things equipment and complex Internet of things environments. The work of the patrol workers is greatly reduced or completely replaced, repeated and boring invalid operation of the patrol workers is reduced, and the dependence on the quality and the enthusiasm of the patrol workers is greatly reduced; the equipment state monitoring data are real-time, reliable and safe, and the safety risks and the data fake or missing risks brought by personnel inspection are reduced.
Furthermore, the embodiment of the application can also monitor the working state of the power equipment in real time, discover the degradation trend of the power equipment at the first time, and provide plentiful time for making maintenance decisions; the state of the power equipment is controlled and predictably maintained, the maintenance interval is prolonged, and technical support is provided for reasonably reducing maintenance cost; avoiding unexpected shutdown until the unplanned shutdown is stopped, and avoiding yield loss caused by unexpected shutdown; and equipment faults and fault reasons are found in time, equipment overhaul guidance is provided for an overhaul party, overhaul time is shortened, and overhaul cost is reduced. The fault part is prejudged in advance, the warehouse is checked in advance, and the problem that standard accessories cannot be found during maintenance, so that people wait for spare parts and waste maintenance time is avoided; or other damage to the equipment in the event of replacement with other emergency spare parts; and evaluating equipment or giving physical reports periodically, optimizing inventory and purchasing plans according to the reports, and optimizing the quantity and structure of spare parts.
In one embodiment, as shown in fig. 3, the voiceprint acquisition terminal of the embodiment of the present application may include a voiceprint collector, a voiceprint processor, and at least one voiceprint sensor, where each voiceprint sensor is connected to the voiceprint collector through a data cable. The voiceprint sensor can acquire original voiceprint data of the power equipment in real time and transmit the original voiceprint data to the voiceprint collector through the data cable. The voiceprint collector can preprocess the original voiceprint data to obtain initial voiceprint data, the initial voiceprint data are transmitted to the intelligent gateway through the network cable according to the communication protocol, and the intelligent gateway sends the initial voiceprint data to the voiceprint analysis cloud platform.
In one embodiment, after receiving the initial voiceprint data of the power device sent by the voiceprint acquisition terminal, the voiceprint analysis cloud platform further includes: preprocessing the initial voiceprint data to update the initial voiceprint data; wherein the preprocessing includes at least one of pre-emphasis, framing, blank rejection, and active frame smoothing.
In one implementation, preprocessing the initial voiceprint data to update the initial voiceprint data includes:
and step A1, compensating the high-frequency component of the initial voiceprint data by using a pre-emphasis algorithm, and increasing the energy of the high-frequency component to compensate the high-frequency energy loss in the transmission process.
And step A2, dividing the compensated initial voiceprint data into a group of short-time frames according to the determined framing interval.
And step A3, eliminating frames with energy smaller than a blank sound threshold value by using a blank sound eliminating algorithm to obtain effective frames.
Step A4, performing smoothing operation on the effective frame by utilizing a Hamming window to obtain a smoothed effective frame, namely updated initial voiceprint data, wherein the smoothed effective frame can be expressed as:
f h (n)＝f(n)m(n)n＝1,2,3,…,N；
where m (N) is a hamming window and N represents the effective frame length.
In one embodiment, performing feature transformation on the initial voiceprint data to obtain time-frequency domain feature data, including: normalizing the initial voiceprint data to obtain normalized data; performing wavelet transformation on the normalized data based on preset super parameters to obtain time frequency spectrums on different wavelet sub-bands; and extracting cepstrum coefficients of the time spectrum on each wavelet sub-band according to the time spectrum on each wavelet sub-band to obtain time-frequency domain characteristic data.
In one implementation, feature transformation is performed on the initial voiceprint data, and the obtained time-frequency domain feature data specifically includes:
and B1, determining super parameters required by preset extraction features, wherein the super parameters comprise a mother wavelet type, a wavelet decomposition level and a cepstrum coefficient number extracted on each wavelet subband.
And step B2, carrying out normalization processing on the preprocessed effective frames, so that the acquisition points of the effective frames meet the normal distribution N (0, 1), and eliminating interference caused by volume difference.
Wherein, f [ n ] is the nth sampling point of the effective frame, and mu and sigma are mean value and standard deviation respectively.
And step B3, performing wavelet transformation on the normalized effective frame by using a Mallat algorithm, wherein the effective frame can be divided into 8 subframes, and the subframes respectively represent the time spectrum of the voiceprint signal on different wavelet subbands. The filtering process can be expressed as:
where h and g are low-pass and high-pass filters, respectively.
Step B4, noise suppression and noise reduction processing is carried out on noise, and a mark D is marked 1 ～D 7 And A 7 For K 1 ～K 8 Wherein, the noise reduction method can be expressed as:
wherein N is i [s]S is the subframe after noise reduction i Is W i Length lambda of (a) i Is the noise threshold.
Step B5, calculating a cepstrum coefficient, wherein the logarithmic power spectrum of the ith wavelet sub-band is defined as:
The ith cepstrum coefficient of the ith wavelet subband is defined as:
wherein, D is the number of cepstrum coefficients extracted from the ith wavelet subband, and I is the number of filters.
And B6, obtaining time-frequency domain characteristic data according to cepstrum coefficients of a time spectrum on each wavelet sub-band, wherein the time-frequency domain characteristic data can be represented by voiceprint characteristic vectors, and specifically comprises the following steps:
and B7, marking the label according to whether the power equipment fails or not in the time period of the extracted voiceprint segment. Considering both normal and fault types, a two-dimensional vector [1,0] is used to represent that the power equipment is normal, and [0,1] is used to represent that the power equipment is faulty. The tag vector is matched with the voiceprint feature vector in association and stored in the cloud platform.
In one embodiment, the voiceprint analysis cloud platform can receive initial voiceprint data of different power devices transmitted by different voiceprint acquisition terminals in different time periods; acquiring a plurality of corresponding records between time-frequency domain characteristic data corresponding to different initial voiceprint data and a fault label, and storing the corresponding records; responding to a demand instruction of a user, selecting at least one corresponding record from a plurality of corresponding records to construct a training sample set, and performing parameter tuning on a preset platform fault identification model through the training sample set. Specifically, the above-described process may include:
In step C1, the voiceprint analysis cloud platform may respond to a user demand instruction, select a specified number of corresponding records, construct a training sample set, and select an infrastructure of a platform fault recognition model, for example, a preset platform fault recognition model may use a deep neural network. Assuming that X is the input dataset and Y is the output dataset, this can be expressed as:
Y＝[y 1 ,…y n ] T ；
wherein m is the number of samples; n is a feature data dimension representing n feature attributes; x is the algorithm input of the deep neural network; y is a label vector, [1,0] represents that the power equipment is normal, and [0,1] represents that the power equipment fails; y is the algorithm target output of the deep neural network.
And C2, the voiceprint analysis cloud platform can respond to an input parameter instruction of a user, and set basic parameters of the deep neural network training, including parameters such as the number of layers of the deep neural network model, the number of units of each layer of the deep neural network model, the learning rate, the training times and the like.
And step C3, training a deep neural network by using an Adam algorithm and a training data set, and modeling a fault recognition model, wherein a model structure diagram of the deep neural network can be seen in FIG. 3. The output of the i-th hidden layer is as follows:
In the formula g i The output result of the ith hidden layer is obtained; omega i ,b i Respectively a weight vector and a threshold vector between the ith group of hidden layers; f is the activation function of the hidden layer; x is input data.
After passing through all hidden layers, the final output layer is reached, and the output layer result can be expressed as:
y＝f′(ω N g N-1 +b N )；
wherein omega is N ，b N Respectively a weight vector and a threshold vector between the last hidden layer and the output layer; f' is the activation function of the output layer; y is the output data.
The method for activating the function is as follows:
the cross entropy is selected as a loss function in deep neural network training, as described below:
in the method, in the process of the application,is the actual output of the neural network; y is the label case of the true result.
Through an Adam algorithm and a training sample set, a trained fault recognition model based on the deep neural network can be obtained through training, and model parameters after tuning, such as weight vectors and threshold vectors among all levels, can be obtained.
According to the embodiment of the application, the tuned model parameters can be stored in the voiceprint analysis cloud platform, an authorized user can conduct management operations such as export, input, modification and the like of the model parameters, and specific model parameter data are issued to the voiceprint acquisition identification terminal according to user requirements.
In one embodiment, as shown in fig. 5, the voiceprint analysis cloud platform can employ a b\s architecture, which can include a voiceprint module and a management module.
The voiceprint module can comprise a platform service sub-module, a data processing sub-module, a model management sub-module and a data interaction sub-module, wherein the platform service sub-module comprises a data cloud storage unit, a cloud computing service unit, a data analysis unit and a visualization unit; the data processing submodule comprises a voiceprint data preprocessing unit, a voiceprint feature extraction unit, a feature sample standard unit and a training sample set construction unit; the model management submodule comprises a training parameter management unit, a training algorithm management unit, an AI algorithm management unit and a fault diagnosis library management unit; the data interaction submodule comprises an acquisition terminal voiceprint data storage unit and a fault diagnosis algorithm model issuing unit.
The management module can comprise a data security sub-module, a data management sub-module and a management configuration sub-module, wherein the data security sub-module comprises an identity verification unit, a user authorization unit, a storage security unit and an identity verification unit; the data management submodule comprises a basic data management unit, a data quality management unit and a data operation and maintenance management unit; the management configuration sub-module comprises a configuration management unit, a task management unit, a log management unit and a monitoring alarm unit. The data security sub-module can manage the user authority, and an authorized user can know the operation condition of the power equipment at any place covered by the Internet at any time.
The voiceprint analysis cloud platform provided by the embodiment of the application can also be communicated with a third party service application to release functions such as operation monitoring, fault alarming, data processing, model training, model release and the like in the service application.
In one embodiment, after receiving the tuned model parameters, the voiceprint acquisition terminal determines a terminal fault identification model built in the voiceprint acquisition terminal according to the tuned model parameters, and performs fault identification on voiceprint information to be identified of the power equipment, which is locally and in real time acquired by the terminal, through the terminal fault identification model, and the method may include:
and D1, the voiceprint acquisition terminal receives the tuned and optimized model parameters issued by the voiceprint analysis cloud platform, and updates parameters of a terminal fault identification model built in the terminal by using the tuned and optimized model parameters.
And D2, collecting voiceprint information of the power equipment in actual operation by the voiceprint collecting terminal, and calculating a feature vector by a local edge calculation method according to the feature transformation to form a feature input sample corresponding to the voiceprint information to be identified.
And D3, performing real-time fault diagnosis on the local power equipment by using the updated terminal fault recognition model and the characteristic input sample, finally outputting a fault diagnosis result, sending the diagnosis result to an echo pattern analysis cloud platform, and timely alarming the fault abnormality of the power equipment by the voiceprint analysis cloud platform.
In one embodiment, as shown in fig. 6, a voiceprint acquisition terminal is arranged around the power equipment such as an air conditioner, a transformer, a power supply, a cabinet and the like, and the voiceprint acquisition terminal acquires voiceprint data in real time and uploads the voiceprint data to a voiceprint analysis cloud platform; preprocessing voiceprint data through a voiceprint analysis cloud platform, wherein the preprocessing comprises pre-emphasis, framing, blank sound rejection, effective frame smoothing and the like; extracting time-frequency domain feature vectors of voiceprint data by using a characteristic transformation method such as wavelet transformation and the like in the voiceprint analysis cloud platform, marking the time-frequency domain feature vectors, and storing the time-frequency domain feature vectors and fault labels corresponding to the time-frequency domain feature vectors in the cloud platform; basic parameters of deep neural network training are input, and a training sample set is constructed by selecting a specific number of time-frequency domain feature vectors from a plurality of stored time-frequency domain feature vectors. Model training is carried out by utilizing a voiceprint analysis cloud platform, the trained model parameters are stored in the cloud platform, management operations such as export, input, modification and the like can be carried out on the trained model parameters, and specific trained model parameter data are issued to a voiceprint acquisition terminal according to user requirements; and the voiceprint acquisition terminal utilizes model parameter data issued based on the voiceprint analysis cloud platform to construct a terminal side fault identification model. And acquiring and monitoring voiceprint data to be identified in real time by using a terminal side fault identification model, locally constructing a test identification sample set, implementing fault diagnosis identification, and identifying potential fault risks of equipment.
The embodiment forms a collection hardware and analysis processing software cloud platform based on the voiceprint of the electric equipment, a voiceprint collection terminal architecture and an arrangement mode are developed in a matched mode, the architecture and the function of a voiceprint analysis cloud platform are developed in a matched mode, the voiceprint collection terminal is used for collecting the voiceprint information of the electric equipment in real time, accumulating operation voiceprint data, recording the voiceprint information data in fault, uploading the voiceprint information data to the voiceprint analysis cloud platform for storage, realizing the functions of voiceprint analysis, visual display and the like, utilizing the voiceprint collection terminal to collect the voiceprint information of the electric equipment in real time, accumulating operation voiceprint data, recording the voiceprint information data in fault, uploading the voiceprint information data to the voiceprint analysis cloud platform for storage, realizing the functions of voiceprint analysis, visual display and the like; integrating a plurality of voiceprint data preprocessing algorithms, voiceprint feature extraction algorithms and artificial intelligent model training algorithms into a voiceprint analysis cloud platform, and realizing iterative updating and rapid migration deployment functions; the voiceprint analysis cloud platform can issue trained model parameters to the voiceprint acquisition terminal according to user requirements, so that localized fault identification of the terminal is realized, voiceprint identification under more complex scenes is facilitated, a solid foundation is provided for adapting to more various Internet of things equipment and complex Internet of things environments, the work of patrol personnel is greatly reduced or completely replaced, and the safety risk, data fake or missing risk brought by personnel patrol are reduced; the solution for integration and integration from the aspects of terminal equipment, communication channels, information models and advanced applications solves the problems of repeated and redundant functions of various terminals and equipment under the condition of meeting various business requirements, reduces the complexity and operation and maintenance cost of the system and improves the usability of the system; the voiceprint analysis cloud platform has the functions of AI intelligent diagnosis, big data expert intelligent diagnosis and the like, an internal algorithm of the cloud platform can be improved and optimized, the functions of iterative optimization and online model updating are achieved, and the expansibility is high.
Based on the same inventive concept, the embodiment of the application also provides a power equipment fault identification method applied to the voiceprint acquisition terminal. The implementation scheme for solving the problem provided by the power equipment fault identification method applied to the voiceprint acquisition terminal is similar to the implementation scheme recorded in the power equipment fault identification method applied to the voiceprint analysis cloud platform, so the specific limitation in the embodiment of the power equipment fault identification method applied to the voiceprint acquisition terminal provided below can be referred to the limitation of the power equipment fault identification method applied to the voiceprint analysis cloud platform hereinabove, and the description is omitted here.
In one embodiment, as shown in fig. 7, there is provided a power equipment fault identification method applied to a voiceprint acquisition terminal, the method including:
step 702, obtaining initial voiceprint data of the power device, and sending the initial voiceprint data to a voiceprint analysis platform.
Step 704, receiving model parameters sent by a voiceprint analysis platform, and determining a terminal fault identification model built in a voiceprint acquisition terminal according to the model parameters, wherein the model parameters are parameters obtained by performing feature transformation on initial voiceprint data by the voiceprint analysis platform, obtaining time-frequency domain feature data, obtaining fault labels corresponding to the time-frequency domain feature data based on the working state of power equipment in a time period corresponding to the initial voiceprint data, and performing parameter tuning on the preset platform fault identification model by utilizing the time-frequency domain feature data and corresponding records of the fault labels.
And step 704, performing fault recognition on the voice print information to be recognized of the power equipment, which is locally and in real time acquired by the terminal, through a terminal fault recognition model.
It should be understood that, although the steps in the flowcharts related to the embodiments described above are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.
Based on the same inventive concept, the embodiment of the application also provides a power equipment fault identification device for realizing the power equipment fault identification method applied to the voiceprint analysis cloud platform. The implementation scheme of the device for solving the problem is similar to the implementation scheme described in the above-mentioned power equipment fault identification method applied to the voiceprint analysis cloud platform, so the specific limitation in the embodiment of the device for identifying power equipment faults provided in one or more embodiments of the device for identifying power equipment faults provided in the following may be referred to the limitation of the power equipment fault identification method applied to the voiceprint analysis cloud platform in the above description, and will not be repeated here.
In one embodiment, as shown in fig. 8, there is provided a power equipment failure recognition apparatus, including: a receive data module 802, an extract annotation module 804, and a tuning parameters module 806, wherein:
and the receiving data module 802 is used for receiving initial voiceprint data of the power equipment sent by the voiceprint acquisition terminal.
The extracting and labeling module 804 is configured to perform feature transformation on the initial voiceprint data to obtain time-frequency domain feature data, and obtain a fault label corresponding to the time-frequency domain feature data based on the working state of the power equipment in a time period corresponding to the initial voiceprint data, where the fault label includes normal power equipment and power equipment fault.
The tuning parameter module 806 is configured to perform parameter tuning on a preset platform fault identification model by using a training sample, obtain a tuned model parameter, and send the tuned model parameter to a voiceprint acquisition terminal, so that the voiceprint acquisition terminal determines a terminal fault identification model built in the voiceprint acquisition terminal according to the tuned model parameter after receiving the tuned model parameter, and performs fault identification on voiceprint information to be identified of the electrical equipment, which is locally and in real time acquired by the terminal, through the terminal fault identification model, where the training sample includes time-frequency domain feature data and a corresponding record of a fault tag.
Based on the same inventive concept, the embodiment of the application also provides a power equipment fault identification device for realizing the power equipment fault identification method applied to the voiceprint acquisition terminal. The implementation scheme of the device for solving the problem is similar to the implementation scheme described in the above-mentioned power equipment fault identification method applied to the voiceprint acquisition terminal, so the specific limitation in the embodiment of the one or more power equipment fault identification devices provided below can be referred to the limitation of the power equipment fault identification method applied to the voiceprint acquisition terminal in the above description, and is not repeated here.
In one embodiment, as shown in fig. 9, there is provided a power equipment failure recognition apparatus, including: an acquisition data module 902, a reception parameters module 904, and a fault identification module 906, wherein:
the acquiring data module 902 is configured to acquire initial voiceprint data of the electrical device, and send the initial voiceprint data to the voiceprint analysis platform.
The receiving parameter module 904 is configured to receive a model parameter sent by the voiceprint analysis platform, and determine a terminal fault identification model built in the voiceprint acquisition terminal according to the model parameter, where the model parameter is a parameter obtained by performing feature transformation on initial voiceprint data by the voiceprint analysis platform, obtaining time-frequency domain feature data, obtaining a fault tag corresponding to the time-frequency domain feature data based on a working state of the power equipment in a time period corresponding to the initial voiceprint data, and performing parameter tuning on a preset platform fault identification model by using the time-frequency domain feature data and a corresponding record of the fault tag, thereby obtaining a tuned parameter.
The fault recognition module 906 is configured to perform fault recognition on voiceprint information to be recognized of the electrical equipment, which is locally collected in real time by using a terminal fault recognition model.
The above-described individual modules in the power equipment failure recognition apparatus may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, a computer device is provided, which may be a server, and the internal structure of which may be as shown in fig. 10. The computer device includes a processor, a memory, an Input/Output interface (I/O) and a communication interface. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface is connected to the system bus through the input/output interface. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is used for storing voice print data of the power device. The input/output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used for communicating with an external terminal through a network connection. The computer program when executed by the processor is used for realizing a power equipment fault identification method applied to the voiceprint acquisition terminal or a power equipment fault identification method applied to the voiceprint analysis cloud platform.
It will be appreciated by those skilled in the art that the structure shown in FIG. 10 is merely a block diagram of some of the structures associated with the present inventive arrangements and is not limiting of the computer device to which the present inventive arrangements may be applied, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having stored therein a computer program, the processor implementing the steps of the method embodiments described above when the computer program is executed.
In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored which, when executed by a processor, implements the steps of the method embodiments described above.
In an embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the steps of the method embodiments described above.
It should be noted that, the user information (including but not limited to user equipment information, user personal information, etc.) and the data (including but not limited to data for analysis, stored data, presented data, etc.) related to the present application are information and data authorized by the user or sufficiently authorized by each party, and the collection, use and processing of the related data need to comply with the related laws and regulations and standards of the related country and region.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magnetic random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (Phase Change Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like. The databases referred to in the embodiments provided herein may include at least one of a relational database and a non-relational database. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processor referred to in the embodiments provided in the present application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, or the like, but is not limited thereto.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The foregoing examples illustrate only a few embodiments of the application and are described in detail herein without thereby limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of the application should be assessed as that of the appended claims.
Claims (10)
1. A method for identifying faults of power equipment, which is characterized by being applied to a voiceprint analysis cloud platform, the method comprising:
receiving initial voiceprint data of the power equipment sent by a voiceprint acquisition terminal;
performing feature transformation on the initial voiceprint data to obtain time-frequency domain feature data, and obtaining fault labels corresponding to the time-frequency domain feature data based on the working state of the power equipment in a time period corresponding to the initial voiceprint data, wherein the fault labels comprise normal power equipment and power equipment faults;
And carrying out parameter tuning on a preset platform fault identification model by using a training sample to obtain tuned model parameters, and transmitting the tuned model parameters to the voiceprint acquisition terminal, so that the voiceprint acquisition terminal determines a terminal fault identification model built in the voiceprint acquisition terminal according to the tuned model parameters after receiving the tuned model parameters, and carries out fault identification on voiceprint information to be identified of the power equipment, which is locally and in real time acquired by the terminal fault identification model, wherein the training sample comprises time-frequency domain feature data and corresponding records of fault labels.
2. The method according to claim 1, wherein the performing feature transformation on the initial voiceprint data to obtain time-frequency domain feature data includes:
normalizing the initial voiceprint data to obtain normalized data;
performing wavelet transformation on the normalized data based on preset super parameters to obtain time frequency spectrums on different wavelet sub-bands;
and extracting a cepstrum coefficient of the time spectrum on each wavelet sub-band according to the time spectrum on each wavelet sub-band to obtain the time-frequency domain characteristic data.
3. The method according to claim 2, further comprising, after receiving the initial voiceprint data from the power device transmitted by the voiceprint acquisition terminal:
preprocessing the initial voiceprint data to update the initial voiceprint data;
wherein the preprocessing includes at least one of pre-emphasis, framing, blank rejection, and active frame smoothing.
4. The method according to claim 1, wherein the method further comprises:
receiving initial voiceprint data of different power equipment in different time periods, wherein the initial voiceprint data are sent by different voiceprint acquisition terminals;
acquiring a plurality of corresponding records between time-frequency domain characteristic data corresponding to different initial voiceprint data and a fault label, and storing the corresponding records;
responding to a demand instruction of a user, selecting at least one corresponding record from the plurality of corresponding records to construct a training sample set, and performing parameter tuning on a preset platform fault recognition model through the training sample set.
5. The utility model provides a power equipment trouble recognition method which is characterized in that is applied to voiceprint acquisition terminal, and the method includes:
acquiring initial voiceprint data of the power equipment, and sending the initial voiceprint data to a voiceprint analysis platform;
Receiving model parameters sent by the voiceprint analysis platform, and determining a terminal fault identification model built in the voiceprint acquisition terminal according to the model parameters, wherein the model parameters are parameters obtained by performing feature transformation on the initial voiceprint data by the voiceprint analysis platform, obtaining time-frequency domain feature data, obtaining fault labels corresponding to the time-frequency domain feature data based on the working state of the power equipment in a time period corresponding to the initial voiceprint data, and performing parameter tuning on a preset platform fault identification model by utilizing the time-frequency domain feature data and the corresponding records of the fault labels to obtain optimized parameters;
and carrying out fault recognition on the voiceprint information to be recognized of the power equipment, which is locally and in real time acquired by the terminal, through the terminal fault recognition model.
6. An electrical equipment fault identification device, the device comprising:
the receiving data module is used for receiving initial voiceprint data of the power equipment sent by the voiceprint acquisition terminal;
the extraction labeling module is used for carrying out feature transformation on the initial voiceprint data to obtain time-frequency domain feature data, and obtaining fault labels corresponding to the time-frequency domain feature data based on the working state of the power equipment in a time period corresponding to the initial voiceprint data, wherein the fault labels comprise normal power equipment and power equipment faults;
The tuning parameter module is used for performing parameter tuning on a preset platform fault identification model by using a training sample to obtain tuned model parameters, and transmitting the tuned model parameters to the voiceprint acquisition terminal, so that the voiceprint acquisition terminal can determine a terminal fault identification model built in the voiceprint acquisition terminal according to the tuned model parameters after receiving the tuned model parameters, and perform fault identification on voiceprint information to be identified of the power equipment, which is locally and in real time acquired by the terminal, through the terminal fault identification model, wherein the training sample comprises time-frequency domain feature data and corresponding records of fault labels.
7. An electrical equipment fault identification device, the device comprising:
the acquisition data module is used for acquiring initial voiceprint data of the power equipment and sending the initial voiceprint data to the voiceprint analysis platform;
the receiving parameter module is used for receiving model parameters sent by the voiceprint analysis platform, determining a terminal fault identification model built in the voiceprint acquisition terminal according to the model parameters, wherein the model parameters are parameters obtained after the voiceprint analysis platform performs feature transformation on the initial voiceprint data to obtain time-frequency domain feature data, and obtains fault labels corresponding to the time-frequency domain feature data based on the working state of the power equipment in a time period corresponding to the initial voiceprint data, and parameter tuning is performed on a preset platform fault identification model by utilizing the time-frequency domain feature data and corresponding records of the fault labels to obtain optimized parameters;
And the fault recognition module is used for carrying out fault recognition on the voice print information to be recognized of the power equipment, which is locally and in real time acquired by the terminal, through the terminal fault recognition model.
8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any one of claims 1 to 5 when the computer program is executed.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 5.
10. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311088522.7A CN117217730A (en) | 2023-08-24 | 2023-08-24 | Power equipment fault identification method, device, equipment, medium and product |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311088522.7A CN117217730A (en) | 2023-08-24 | 2023-08-24 | Power equipment fault identification method, device, equipment, medium and product |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117217730A true CN117217730A (en) | 2023-12-12 |
Family
ID=89039885
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311088522.7A Pending CN117217730A (en) | 2023-08-24 | 2023-08-24 | Power equipment fault identification method, device, equipment, medium and product |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117217730A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117894317A (en) * | 2024-03-14 | 2024-04-16 | 沈阳智帮电气设备有限公司 | Box-type transformer on-line monitoring method and system based on voiceprint analysis |
| CN118100444A (en) * | 2024-04-19 | 2024-05-28 | 国网四川省电力公司广安供电公司 | Low-voltage distribution network fault mode analysis method |
-
2023
- 2023-08-24 CN CN202311088522.7A patent/CN117217730A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117894317A (en) * | 2024-03-14 | 2024-04-16 | 沈阳智帮电气设备有限公司 | Box-type transformer on-line monitoring method and system based on voiceprint analysis |
| CN117894317B (en) * | 2024-03-14 | 2024-05-24 | 沈阳智帮电气设备有限公司 | Box-type transformer on-line monitoring method and system based on voiceprint analysis |
| CN118100444A (en) * | 2024-04-19 | 2024-05-28 | 国网四川省电力公司广安供电公司 | Low-voltage distribution network fault mode analysis method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117217730A (en) | Power equipment fault identification method, device, equipment, medium and product | |
| Tcherniak et al. | Active vibration-based structural health monitoring system for wind turbine blade: Demonstration on an operating Vestas V27 wind turbine | |
| CN111325095B (en) | Intelligent detection method and system for equipment health state based on acoustic wave signals | |
| CN111177095A (en) | Log analysis method and device, computer equipment and storage medium | |
| CN108008332A (en) | A kind of new energy Remote testing device method for diagnosing faults based on data mining | |
| US20100030521A1 (en) | Method for analyzing and classifying process data | |
| CN110398647B (en) | Transformer state monitoring method | |
| CN114973155B (en) | Intelligent monitoring, analyzing and managing system based on AI image recognition behaviors | |
| CN113670434B (en) | Method and device for identifying sound abnormality of substation equipment and computer equipment | |
| CN115640915A (en) | Intelligent gas pipe network compressor safety management method and Internet of things system | |
| CN113112038B (en) | Intelligent monitoring and diagnostic analysis system, device, electronic equipment and storage medium | |
| CN113257249A (en) | Power equipment fault diagnosis method, device and equipment based on voiceprint recognition | |
| CN113612625A (en) | Network fault positioning method and device | |
| CN112200238A (en) | Hard rock tension-shear fracture identification method and device based on sound characteristics | |
| CN115376526A (en) | Power equipment fault detection method and system based on voiceprint recognition | |
| Jarwar et al. | NEAT: A resilient deep representational learning for fault detection using acoustic signals in IIoT environment | |
| CN113497725A (en) | Alarm monitoring method, alarm monitoring system, computer readable storage medium and electronic equipment | |
| CN115877068A (en) | Voltage sag propagation track identification method of regional power grid based on deep learning | |
| CN113707175B (en) | Acoustic event detection system based on feature decomposition classifier and adaptive post-processing | |
| CN117292713A (en) | Transformer voiceprint monitoring method and system based on end Bian Yun cooperation | |
| CN110197289B (en) | Energy-saving equipment management system based on big data | |
| CN114997749B (en) | Intelligent scheduling method and system for power personnel | |
| CN116129935A (en) | Voiceprint data acquisition system for mining transformer | |
| CN114760194A (en) | Fault studying and judging method for Internet of things equipment | |
| CN113886592A (en) | Quality detection method for operation and maintenance data of power information communication system |
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 |