Fault detection method and device, electronic equipment and readable storage medium

Abstract

Embodiments of the present application provide a fault detection method and device, electronic equipment and a readable storage medium, the method comprising: collecting audio data generated when a to-be-detected device fails; using a pre-constructed first classification model to perform classification processing on the audio data to obtain a first fault category of the to-be-detected device and a confidence corresponding to the first fault category; if the confidence is less than a first preset threshold, using a pre-constructed second classification model to perform classification processing on the audio data to obtain a second fault category of the to-be-detected device and a similarity between the audio data and a target audio template; and determining a fault detection result of the to-be-detected device according to the second fault category. Through the above method, the embodiments of the present application realize accurate detection of frequent faults and non-frequent faults of a device in a complex production environment, which helps to improve production efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of acoustics, and in particular to a fault detection method and device, electronic equipment and a readable storage medium. BACKGROUND

[0002] In industrial systems, the sound generation mechanism and propagation law are complex, and the environmental conditions are diverse. Therefore, different faults of different devices usually have different acoustic characteristics. Therefore, it is a common method to determine the specific fault by the sound of the fault.

[0003] With the development of big data, big data has also been applied to fault detection based on acoustics.

[0004] However, big data algorithms require a large amount of voiceprint data, especially acoustic data when the device is faulty. The internal noise generation mechanism of industrial equipment, especially large and complex equipment, is complex, and the transmission law of sound from the internal structure to the external sensor is complex, penetrating through various media. In addition, the background noise of the device and the environment is complex and diverse, resulting in uneven training samples of the big data model, and thus the accuracy of fault detection classification is low. SUMMARY

[0005] The embodiments of the present application provide a fault detection method, device, electronic equipment and readable storage medium, which can realize accurate fault detection based on audio data in a complex environment.

[0006] In a first aspect, the embodiments of the present application provide a fault detection method, which comprises:

[0007] Collecting audio data generated when a device to be detected fails;

[0008] Using a pre-constructed first classification model to classify the audio data, obtaining a first fault category of the device to be detected and a confidence level corresponding to the first fault category;

[0009] If the confidence level is less than a first preset threshold, a pre-constructed second classification model is used to classify the audio data, obtaining a second fault category of the device to be detected and a similarity between the audio data and a target audio template; the similarity between the audio data and the target audio model is greater than the similarity between the audio data and other audio templates in the second classification model; the second fault category is the fault category corresponding to the target audio template;

[0010] According to the second fault category, a fault detection result of the device to be detected is determined.

[0011] Optionally, the step of determining the fault detection result of the device to be detected according to the second fault category comprises:

[0012] If the similarity is greater than or equal to a second preset threshold, it is determined that the failure category of the to-be-detected device is the second failure category.

[0013] If the similarity is less than the second preset threshold, the to-be-detected device is subjected to failure detection processing, and a target detection result of the to-be-detected device is obtained.

[0014] Optionally, the target detection result includes a third failure category of the to-be-detected device; and the method further includes:

[0015] generating a category label of the audio data according to the third failure category;

[0016] updating the second classification model according to the audio data and the category label of the audio data, so as to take the audio data as an audio template corresponding to the category label in the second classification model.

[0017] Before the classification processing of the audio data by using the pre-constructed second classification model, the method further includes:

[0018] obtaining a first audio sample, the first audio sample including first audio data generated when a target device occurs an infrequent failure and a first category label corresponding to the first audio data;

[0019] taking the first audio data as an audio template corresponding to the second category label in a second classification model, and constructing the second classification model; the second classification model is used to calculate a similarity between input audio data and each audio template, and determine a failure category corresponding to the input audio data based on the similarity.

[0020] Optionally, the classification processing of the audio data by using the pre-constructed second classification model, obtaining a second failure category of the to-be-detected device and a similarity between the audio data and a target audio template, includes:

[0021] calculating the similarity between the audio data and each audio template by using the second failure classification model;

[0022] sorting the audio templates according to the similarity between the audio data and each audio template, and obtaining a sorting result;

[0023] determining a target audio template corresponding to the audio data according to the sorting result;

[0024] obtaining a second category label corresponding to the target audio template, and determining the second category label corresponding to the target audio template as the second failure category of the to-be-detected device.

[0025] Optionally, before the audio data is classified by using the pre-constructed first classification model, the method further comprises:

[0026] obtaining a second audio sample, the second audio sample comprising second audio data generated when the target device frequently malfunctions and a second category label corresponding to the second audio data;

[0027] extracting acoustic features of the audio data in the second audio sample to obtain a second training sample;

[0028] iteratively training a neural network model by using the second training sample until a training termination condition is met to obtain a first classification model.

[0029] In a second aspect, an embodiment of the present application provides a fault detection device, the device comprising:

[0030] a data acquisition module configured to acquire audio data generated when a device to be detected malfunctions;

[0031] a first classification module configured to classify the audio data by using a pre-constructed first classification model to obtain a first fault category of the device to be detected and a confidence degree corresponding to the first fault category;

[0032] a second classification module configured to, if the confidence degree is less than a first preset threshold, classify the audio data by using a pre-constructed second classification model to obtain a second fault category of the device to be detected and a similarity between the audio data and a target audio template; the similarity between the audio data and the target audio template is greater than similarities between the audio data and other audio templates in the second classification model; and the second fault category is a fault category corresponding to the target audio template.

[0033] a fault detection result determination module configured to determine a fault detection result of the device to be detected according to the second fault category.

[0034] In a third aspect, an embodiment of the present application provides an electronic device, comprising a processor and a memory, the processor executing a computer program stored in the memory to implement the fault detection method of the first aspect.

[0035] In a fourth aspect, an embodiment of the present application provides a readable storage medium, the readable storage medium storing computer instructions, the computer instructions being executed by a processor to implement the fault detection method of the first aspect.

[0036] Embodiments of the present application have the following advantages:

[0037] In summary, the present application provides an embodiment of a fault detection method. In the embodiment of the present application, after collecting audio data generated when the to-be-detected device fails, first, a first classification model constructed in advance is used to classify the audio data, to obtain a first fault category of the to-be-detected device and a confidence corresponding to the first fault category. If the confidence is less than a first preset threshold, it indicates that the fault category of the to-be-detected device is not a frequent fault. Then, a second classification model constructed in advance is used to classify the audio data, to obtain a second fault category of the to-be-detected device and a similarity between the audio data and a target audio template. Finally, the fault detection result of the to-be-detected device is determined according to the second fault category. Through the above method, the embodiment of the present application realizes accurate detection of frequent faults and non-frequent faults of the device in a complex production environment, which helps to improve production efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the description of the embodiments of the present application. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor under the premise of the drawings.

[0039] Figure 1 A flow chart of an embodiment of a fault detection method of the present application is shown;

[0040] Figure 2 A structural block diagram of an embodiment of a fault detection device of the present application is shown;

[0041] Figure 3 A structural schematic diagram of an electronic device provided by the embodiment of the present application is shown. DETAILED DESCRIPTION

[0042] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application. And it should be noted that the acquisition of various data related processes in the embodiments of the present application is under the premise of complying with the corresponding data protection regulations and policies of the place, and with the authorization given by the corresponding device owner.

[0043] Referring to Figure 1 , a flow chart of an embodiment of a fault detection method of the present application is shown, which can specifically include the following steps:

[0044] Step 101, collect audio data generated when the to-be-detected equipment fails.

[0045] In an industrial system, the inside of a production equipment is subjected to electric, magnetic, mechanical and other stresses, which will generate vibrations, and mechanical waves formed by the vibrations are transmitted to the shell through the medium, and the sound source signal can be captured by a sensor device. The signal contains a large amount of time-frequency domain feature information, like a fingerprint. When the production equipment is abnormal, the acoustic fingerprint, simply referred to as "voiceprint", will change, which can be used as a main characteristic parameter for diagnosing equipment defects and failures.

[0046] Therefore, in the embodiment of the present application, when a failure occurs, the voiceprint to be detected will change, and the audio data generated when the to-be-detected equipment fails can be collected by a sensor or other sound sensing device.

[0047] Step 102, using a pre-constructed first classification model, classifying the audio data to obtain a first fault category of the to-be-detected equipment and a confidence degree corresponding to the first fault category.

[0048] Frequent failure refers to a failure that occurs frequently for equipment. For example, when a transformer fails, the transformer itself will emit a howling sound; when the transformer is overloaded or the coil current is large, the transformer itself will produce a loud and deep sound; when the transformer core is grounded and the wire breaks, the transformer itself will produce a crackling sound.

[0049] For frequent failures, the sensor is more likely to collect a large number of audio data samples, and the classification effect of the classification model based on big data will be better. Therefore, in the embodiment of the present application, the first classification model can be a classification model based on a neural network, which is used for classifying the frequent failures of the to-be-detected equipment, and can include a convolutional neural network (CNN), a time delay neural network (TDNN), etc., which is not limited in the present application.

[0050] In the embodiment of the present application, the neural network classification model is trained by collecting a large number of audio data samples to construct the first classification model, and then the audio data of the to-be-detected equipment collected in step 101 is input into the first classification model, so as to output the first fault category of the to-be-detected equipment and the confidence degree corresponding to the first fault category. The confidence degree refers to the probability that the failure of the to-be-detected equipment obtained by the first classification model is the first fault category. For example, the first fault category is short circuit, and the confidence degree corresponding to the short circuit is 0.8.

[0051] In an optional embodiment of the present application, before the step 102 of classifying the audio data by using the pre-constructed first classification model, the method further comprises:

[0052] The step S100 comprises: acquiring a second audio sample, the second audio sample comprising second audio data generated when the target device frequently malfunctions and a second category label corresponding to the second audio data.

[0053] The step S101 comprises: extracting acoustic features of the audio data in the second audio sample to obtain a second training sample.

[0054] The step S102 comprises: iteratively training a neural network model by using the second training sample until a training termination condition is met to obtain a first classification model.

[0055] The target device refers to a collection object when the frequently malfunctioning audio data, i.e., the second audio data, is collected. The acoustic features can include fbank features, Mel-scale Frequency Cepstral Coefficients (MFCC), etc., which are not limited in the present application. The second category label can be a category of the malfunction, such as short circuit, or a sound category corresponding to the malfunction category, such as a howling sound corresponding to short circuit.

[0056] In order to improve the efficiency of model training, the second category label can be labeled by using the file name of the second audio data. As an example, the file name of the second audio data can be PaPaSheng1.wav, duanlu.wav, HuXiaoSheng 3.wav, wherein duanlu and huxiao3 can represent that the malfunction category of the transformer is short circuit. Of course, the above file names are only used to explain the second category label and have no limiting effect.

[0057] By iteratively training the neural network model by collecting a large amount of frequently malfunctioning audio data, the classification accuracy of the neural network model can be effectively improved, thereby improving the classification accuracy of the frequently malfunctioning in the complex industrial production environment.

[0058] The step 103 comprises: if the confidence is less than a first preset threshold, classifying the audio data by using a pre-constructed second classification model to obtain a second malfunction category of the device to be detected and a similarity between the audio data and a target audio template; the similarity between the audio data and the target audio model is greater than the similarity between the audio data and other audio templates in the second classification model; and the second malfunction category is a malfunction category corresponding to the target audio template.

[0059] The non-frequent fault refers to a fault with a low occurrence frequency for the equipment. For example, when there is a stain in the transformer bushing, the enamel has damage or a crack to cause corona flashover, the transformer will produce a crackling sound. Since the crackling sound is a fault caused by device damage, it is a non-frequent fault. Of course, in actual industrial production, the non-frequent fault can be divided according to specific conditions, which is not limited in the present application.

[0060] The first preset threshold refers to a threshold for judging the confidence of the first fault category. If the confidence corresponding to the first fault category is less than the first preset threshold, it can be considered that the fault category of the to-be-detected equipment is not the first fault category. The first preset threshold can be set according to specific conditions in industrial production, which is not limited in the present application.

[0061] The second classification model can be a classification model based on dynamic time warping (DTW). In the case that the confidence corresponding to the first fault category is less than the first preset threshold, i.e., the first fault category is inaccurate, the second classification model can be used to classify the fault of the to-be-detected equipment again. Specifically, the DTW can use a dynamic programming (DP) method to calculate the time warping, and can calculate the similarity of two time series, especially suitable for time series with different lengths and different rhythms, such as audio sequences of different people reading the same word. In the embodiments of the present application, the similarity between the audio data obtained in step 101 and the audio templates in the first classification model can be calculated respectively, and the fault category corresponding to the audio template with the highest similarity to the audio data, i.e., the target fault template, is taken as the fault classification result of the to-be-detected equipment, i.e., the second fault category.

[0062] For example, the audio data and the audio template can both be time series of volume decibels, so in the case of the same time scale, the volume difference at the same time point can be calculated, and the volume difference at each time point can be integrated by weighting, square root, etc. to obtain the integrated volume difference between the two. Taking the integrated volume difference as the standard, the higher the integrated volume difference, the lower the similarity between the audio data and the audio template. Of course, the above-mentioned time series of volume decibels are only used to explain the audio data and the audio template, and have no limiting effect on the audio data and the audio template. They can be calculated according to actual conditions.

[0063] In the case that the sample data of the non-frequent fault is difficult to collect, the second classification model based on DTW can be constructed using limited sample audio data of the non-frequent fault, which can better realize accurate classification of the non-frequent fault in a complex environment.

[0064] In an optional embodiment of the present application, before the step 103 of classifying the audio data by using the pre-constructed second classification model, the method further comprises:

[0065] The step S200 acquires a first audio sample, which includes first audio data generated when the target device has a non-frequent fault and a first category label corresponding to the first audio data.

[0066] The step S201 takes the first audio data as an audio template corresponding to the second category label in the second classification model, and constructs a second classification model; the second classification model is used to calculate the similarity between input audio data and each audio template, and determine the fault category corresponding to the input audio data based on the similarity.

[0067] In order to construct the second classification model and enrich the audio templates in the second classification model, similar to the collection of the second audio sample, the sound, i.e. audio data, emitted by the device when the device has a non-frequent fault can be collected, and the audio data is labeled by category, so as to calculate the similarity between the audio data of the device to be detected and the audio template by using the second classification model.

[0068] Further, based on the steps S200-S201, in an optional embodiment of the present application, the step 103 of classifying the audio data by using the pre-constructed second classification model to obtain the second fault category of the device to be detected and the similarity between the audio data and the target audio template can include:

[0069] The step S300 calculates the similarity between the audio data and each audio template by using the second fault classification model.

[0070] The step S301 sorts the audio templates according to the similarity between the audio data and each audio template to obtain a sorting result.

[0071] The step S302 determines a target audio template corresponding to the audio data according to the sorting result.

[0072] The step S303 acquires a second category label corresponding to the target audio template, and determines the second category label corresponding to the target audio template as the second fault category of the device to be detected.

[0073] After the second classification model is constructed, when the audio data of the to-be-detected device is classified by using the second classification model, the second classification model can be used to calculate the similarity of the audio data of the to-be-detected device and each audio template, and the category corresponding to the audio template with the highest similarity is selected as the second fault category of the to-be-detected device, so that the non-occasional fault classification of the to-be-detected device is realized.

[0074] Step 104, determining the fault detection result of the to-be-detected device according to the second fault category.

[0075] Similar to the first fault category, if the similarity of the second fault category and the target audio template is too low, it can be considered that the second fault category is not necessarily the fault detection result of the to-be-detected device.

[0076] Therefore, in an optional embodiment of the present application, the step 104 of determining the fault detection result of the to-be-detected device according to the second fault category can include:

[0077] Step S400, if the similarity is greater than or equal to a second preset threshold, determining that the fault category of the to-be-detected device is the second fault category;

[0078] Step S401, if the similarity is less than the second preset threshold, performing fault detection processing on the to-be-detected device to obtain a target detection result of the to-be-detected device.

[0079] The second preset threshold refers to a threshold for judging the similarity of the second fault category. If the similarity corresponding to the second fault category is greater than or equal to the second preset threshold, it can be considered that the fault category of the to-be-detected device is the second fault category. The second preset threshold can be set according to the specific situation in industrial production, which is not limited in the present application.

[0080] If the similarity corresponding to the second fault category is less than the second preset threshold, it can be considered that the similarity of the second fault category and the target audio template is too low. The to-be-detected device can be physically detected by a physical detection method, so as to obtain the fault detection result of the to-be-detected device, i.e., the target detection result.

[0081] On the basis of the method in steps S400-S401, the target detection result includes a third fault category of the to-be-detected device. The method of the present application can further include:

[0082] Step S500, generating a category label of the audio data according to the third fault category;

[0083] Step S501, updating the second classification model according to the audio data and the category label of the audio data, so as to take the audio data as an audio template corresponding to the category label in the second classification model.

[0084] In order to improve the accuracy of the second classification model and the identification range of fault identification, the audio data of the to-be-detected equipment can be labeled according to the third fault category, and the audio data and the category label of the audio data are taken as a set of audio templates to update the second classification model.

[0085] In summary, the embodiment of the fault detection method is provided. In the embodiment of the present application, after collecting the audio data generated when the to-be-detected equipment fails, first, the first classification model is used to classify the audio data, and the first fault category of the to-be-detected equipment and the confidence corresponding to the first fault category are obtained. If the confidence is less than the first preset threshold, it indicates that the fault category of the to-be-detected equipment is not a frequent fault. Then, the second classification model is used to classify the audio data, and the second fault category of the to-be-detected equipment and the similarity between the audio data and the target audio template are obtained. Finally, the fault detection result of the to-be-detected equipment is determined according to the second fault category. Through the above method, the embodiment of the present application realizes accurate detection of frequent faults and non-frequent faults of the equipment in a complex production environment, which helps to improve the production efficiency.

[0086] Referring to Figure 2 , a structural block diagram of an embodiment of a fault detection device of the present application is shown, and the device 200 can include:

[0087] The data acquisition module 201 is configured to acquire audio data generated when the to-be-detected equipment fails.

[0088] The first classification module 202 is configured to use a pre-constructed first classification model to classify the audio data, and obtain a first fault category of the to-be-detected equipment and a confidence corresponding to the first fault category.

[0089] The second classification module 203 is configured to use a pre-constructed second classification model to classify the audio data if the confidence is less than a first preset threshold, and obtain a second fault category of the to-be-detected equipment and a similarity between the audio data and a target audio template. The similarity between the audio data and the target audio model is greater than the similarity between the audio data and other audio templates in the second classification model. The second fault category is a fault category corresponding to the target audio template.

[0090] The fault detection result determination module 204 is configured to determine the fault detection result of the to-be-detected device according to the second fault category.

[0091] Optionally, the fault detection result determination module can include:

[0092] The first determination sub-module is configured to determine that the fault category of the to-be-detected device is the second fault category if the similarity is greater than or equal to a second preset threshold.

[0093] The second determination sub-module is configured to perform fault detection processing on the to-be-detected device to obtain a target detection result of the to-be-detected device if the similarity is less than the second preset threshold.

[0094] Optionally, the target detection result includes a third fault category of the to-be-detected device.

[0095] The apparatus can further include:

[0096] The category label generation module is configured to generate a category label of the audio data according to the third fault category.

[0097] The audio template generation module is configured to update the second classification model according to the audio data and the category label of the audio data, so as to take the audio data as an audio template corresponding to the category label in the second classification model.

[0098] Optionally, the apparatus can further include:

[0099] The first audio data acquisition module is configured to acquire a first audio sample before the audio data is classified by using the pre-constructed second classification model, the first audio sample including first audio data generated when a target device has a non-frequent fault and a first category label corresponding to the first audio data.

[0100] The second classification model construction module is configured to take the first audio data as an audio template corresponding to the second category label in a second classification model, and to construct the second classification model; the second classification model is configured to calculate the similarity between input audio data and each audio template, and to determine the fault category corresponding to the input audio data based on the similarity.

[0101] Optionally, the second classification model can include:

[0102] The similarity calculation sub-module is configured to calculate the similarity between the audio data and each audio template by using the second fault classification model.

[0103] The sorting submodule is configured to sort the audio templates according to the similarity between the audio data and each audio template, to obtain a sorting result.

[0104] The target audio template determination submodule is configured to determine a target audio template corresponding to the audio data according to the sorting result.

[0105] The second fault category determination submodule is configured to obtain a second category label corresponding to the target audio template, and determine the second category label corresponding to the target audio template as a second fault category of the to-be-detected device.

[0106] Optionally, the apparatus can further include:

[0107] The second audio sample acquisition module is configured to acquire a second audio sample, the second audio sample including second audio data generated when a target device frequently malfunctions and a second category label corresponding to the second audio data.

[0108] The feature extraction module is configured to extract an acoustic feature of the audio data in the second audio sample, to obtain a second training sample.

[0109] The first classification model training module is configured to iteratively train a neural network model by using the second training sample until a training termination condition is met, to obtain a first classification model.

[0110] Referring to Figure 3 The electronic device 600 can include one or more of the following components: a processing component 602, a memory 604, a power supply component 606, a multimedia component 608, an audio component 610, an input / output (I / O) interface 612, a sensor component 614, and a communication component 616.

[0111] The processing component 602 generally controls the overall operation of the electronic device 600, such as operations associated with displaying, making phone calls, data communications, camera operations, and recording operations. The processing component 602 can include one or more processors 620 to execute instructions to complete all or part of the steps of the fault detection method described above. In addition, the processing component 602 can include one or more modules to facilitate interaction between the processing component 602 and other components. For example, the processing component 602 can include a multimedia module to facilitate interaction between the multimedia component 608 and the processing component 602.

[0112] The memory 604 is configured to store various types of data to support the operation of the electronic device 600. Examples of such data include instructions for any application or method operating on the electronic device 600, contact data, phonebook data, messages, pictures, videos, etc. The memory 604 can be implemented by any type of volatile or nonvolatile memory, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disc, or optical disc.

[0113] The power supply component 606 supplies power for various components of the electronic device 600. The power supply component 606 can include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the electronic device 600.

[0114] The multimedia component 608 includes a screen providing an output interface between the electronic device 600 and a user. In some embodiments, the screen can include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen can be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touch, swiping, and gestures on the touch panel. The touch sensors can not only sense a boundary of a touching or swiping action, but also detect duration and pressure related to the touching or swiping action. In some embodiments, the multimedia component 608 includes a front camera and / or a rear camera. The front and rear cameras can receive external multimedia data when the electronic device 600 is in an operation mode, such as a photographing mode or a video mode. Each of the front and rear cameras can be a fixed optical lens system or have a focal length and optical zoom capability.

[0115] The audio component 610 is configured to output and / or input audio signals. For example, the audio component 610 includes a microphone (MIC) configured to receive external audio signals when the electronic device 600 is in an operation mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals can be further stored in the memory 604 or transmitted via the communication component 616. In some embodiments, the audio component 610 also includes a speaker for outputting audio signals.

[0116] The I / O interface 612 provides an interface between the processing component 602 and peripheral interface modules, which can be a keypad, a click wheel, buttons, etc. The buttons can include, but are not limited to, a home button, a volume button, a start button, and a lock button.

[0117] The sensor component 614 includes one or more sensors for providing status assessments for various aspects of the electronic device 600. For example, the sensor component 614 can detect an open / closed position of the electronic device 600, relative positioning of components, such as a display and keypad of the device 600, a change in position of the electronic device 600 or a component of the electronic device 600, the presence or absence of user contact with the electronic device 600, the orientation or acceleration / deceleration / g-force and a temperature change of the electronic device 600. The sensor component 614 can include an optical sensor for detecting ambient light, a proximity sensor for detecting nearby objects without any physical touch, a CMOS or CCD image sensor for use in imaging applications, or an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor in some embodiments.

[0118] The communication component 616 is configured to facilitate wired or wireless communication between the electronic device 600 and other devices. The electronic device 600 can access a wireless network based on a communication standard, such as WiFi, 2G, or 3G, or a combination thereof. In an example embodiment, the communication component 616 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In an example embodiment, the communication component 616 further includes a Near Field Communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on Radio Frequency Identification (RFID) techniques, infrared data association (IrDA) techniques, ultra-wideband (UWB) techniques, Bluetooth (BT) techniques, and other techniques.

[0119] In an example embodiment, the electronic device 600 can be implemented using one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, micro-controllers, microprocessors, or other electronic elements for performing the above-described fault detection method.

[0120] In an example embodiment, a non-transitory computer-readable storage medium including instructions, such as the memory 604 including instructions, is also provided, which can be executed by the processor 620 of the electronic device 600 to implement the above-described fault detection method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disc, and an optical data storage device, etc.

[0121] A non-transitory computer readable storage medium, when instructions in the storage medium are executed by a processor of an electronic device (a server or a terminal), enable the processor to perform the fault detection method described above.

[0122] The various embodiments are described in the specification by way of progressive progression, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be mutually referred to. For the apparatus disclosed by the embodiments, since it corresponds to the method disclosed by the embodiments, the description is relatively simple, and the relevant parts can be referred to the method part.

[0123] Finally, it should be noted that, in this document, the relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitations, the element defined by the statement "including a" does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0124] The above describes in detail the fault detection method, device, electronic device and storage medium provided by the present application. The principles and implementation manners of the present application are described by applying specific examples. The above embodiment description is only used to help understand the method of the present application and its core idea; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation manner and application range will be changed; in summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A fault detection method characterized by, The method comprises: collecting audio data generated when a device to be detected fails; obtaining a second audio sample, the second audio sample comprising second audio data generated when a target device frequently fails and a second category label corresponding to the second audio data; extracting acoustic features of the audio data in the second audio sample to obtain a second training sample; iteratively training a neural network model using the second training sample until a training termination condition is met to obtain a first classification model; classifying the audio data using the pre-constructed first classification model to obtain a first failure category of the device to be detected and a confidence level corresponding to the first failure category; if the confidence level is less than a first preset threshold, classifying the audio data using a pre-constructed second classification model to obtain a second failure category of the device to be detected and a similarity between the audio data and a target audio template; the similarity between the audio data and the target audio template is greater than the similarity between the audio data and other audio templates in the second classification model; the second failure category is a failure category corresponding to the target audio template; determining a failure detection result of the device to be detected according to the second failure category; before classifying the audio data using the pre-constructed second classification model, the method further comprises: obtaining a first audio sample, the first audio sample comprising first audio data generated when a target device fails non-frequently and a first category label corresponding to the first audio data; using the first audio data as an audio template corresponding to a second category label in a second classification model, and constructing a second classification model; the second classification model is used to calculate the similarity between input audio data and each audio template, and determine a failure category corresponding to the input audio data based on the similarity.

2. The method of claim 1, wherein, The determination of the failure detection result of the device to be detected according to the second failure category comprises: if the similarity is greater than or equal to a second preset threshold, determining that the failure category of the device to be detected is the second failure category; if the similarity is less than the second preset threshold, performing failure detection processing on the device to be detected to obtain a target detection result of the device to be detected.

3. The method of claim 2, wherein, The target detection result comprises a third failure category of the device to be detected; the method further comprises: generating a category label of the audio data according to the third failure category; updating the second classification model according to the audio data and the category label of the audio data, so as to use the audio data as an audio template corresponding to the category label in the second classification model.

4. The method of claim 1, wherein, The classification of the audio data using the pre-constructed second classification model to obtain the second failure category of the device to be detected and the similarity between the audio data and the target audio template comprises: calculating the similarity between the audio data and each audio template using the second classification model; sorting the audio templates according to the similarity between the audio data and each audio template to obtain a sorting result; determine a target audio template corresponding to the audio data according to the ranking result; obtain a second category label corresponding to the target audio template, and determine the second category label corresponding to the target audio template as a second fault category of the to-be-detected device.

5. A fault detection apparatus characterized by comprising: The apparatus comprises: a data collection module configured to collect audio data generated when a to-be-detected device fails; a first classification module configured to perform classification processing on the audio data by using a pre-constructed first classification model, to obtain a first fault category of the to-be-detected device and a confidence degree corresponding to the first fault category; a second classification module configured to, if the confidence degree is less than a first preset threshold, perform classification processing on the audio data by using a pre-constructed second classification model, to obtain a second fault category of the to-be-detected device and a similarity between the audio data and a target audio template; the similarity between the audio data and the target audio template is greater than similarities between the audio data and other audio templates in the second classification model; and the second fault category is a fault category corresponding to the target audio template; a fault detection result determination module configured to determine a fault detection result of the to-be-detected device according to the second fault category. The apparatus further comprises: a second audio sample collection module configured to collect a second audio sample, the second audio sample comprising second audio data generated when a target device frequently fails and a second category label corresponding to the second audio data; a feature extraction module configured to extract acoustic features of audio data in the second audio sample, to obtain a second training sample; a first classification model training module configured to iteratively train a neural network model by using the second training sample until a training termination condition is met, to obtain a first classification model; a first audio data collection module configured to, before performing classification processing on the audio data by using the pre-constructed second classification model, collect a first audio sample, the first audio sample comprising first audio data generated when a target device does not frequently fail and a first category label corresponding to the first audio data; a second classification model construction module configured to use the first audio data as an audio template corresponding to the second category label in a second classification model, and to construct the second classification model; the second classification model is configured to calculate similarities between input audio data and each audio template, and to determine a fault category corresponding to the input audio data based on the similarities.

6. The apparatus of claim 5, wherein, The fault detection result determination module comprises: a first determination submodule configured to, if the similarity is greater than or equal to a second preset threshold, determine that a fault category of the to-be-detected device is the second fault category; a second determination submodule configured to, if the similarity is less than the second preset threshold, perform fault detection processing on the to-be-detected device, to obtain a target detection result of the to-be-detected device.

7. An electronic device, comprising: The apparatus comprises: a processor and a memory, the processor executes a computer program stored in the memory, and implements the fault detection method in any one of claims 1 to 4.

8. A readable storage medium, characterized by, When the instructions in the storage medium are executed by the processor of the device, the device is enabled to perform the fault detection method as claimed in any one of the method claims 1 to 4.

Images