Method and computing device for determining the operating state and fault type of a water pump

Abstract

The present invention discloses a method and a computing device for determining the operating state and fault type of a water pump. The method includes: obtaining audio data when the water pump is operating, where the audio data includes the respective audio data of the water pump and the water in the water pump, and the audio data of the water pump is the first audio data, and the audio data of the water is the second audio data; extracting the first audio data from the audio data by using a trained voiceprint extraction model; comparing the first audio data with the audio sample data stored in the voiceprint database to determine the operating state of the water pump; and when it is determined that the operating state of the water pump is a fault, determining the fault type of the fault. By automatically collecting the audio when the water pump is operating, performing noise reduction processing on the audio and analyzing it, the present invention can determine the operating state and fault type of the water pump, greatly improving the recognition efficiency, comprehensiveness and heredity of water pump faults and their types.

Description

Technical Field

[0001] The present invention relates to the field of water pumps, and more particularly to a method and a computing device for determining the operating state and fault type of a water pump. Background Art

[0002] Water pumps are widely used in modern industry. When a water pump fails, the location and type of the fault are often judged by manual inspection. However, some faults cannot be manually inspected when the water pump is running, resulting in a long cycle for fault discovery. In addition, faults such as cavitation of the water pump cannot be detected by manual inspection, making it difficult for the manual inspection method to cover all times and all aspects, and there are risks of misdetection and undetected faults. Moreover, manual inspection for fault location relies on human ear for sound discrimination, but the human ear sound discrimination technology depends on expert experience transmission and long-term learning accumulation, with a high learning cost.

[0003] Therefore, a new method and a computing device for determining the operating state and fault type of a water pump are needed to solve the above problems. Summary of the Invention

[0004] A series of simplified concepts are introduced in the Summary of the Invention section, which will be further described in detail in the Detailed Description section. The Summary of the Invention section of the present invention does not mean to attempt to define the key features and essential technical features of the claimed technical solution, nor does it mean to attempt to determine the protection scope of the claimed technical solution.

[0005] According to an embodiment of the present invention, there is provided a method for determining the operating state and fault type of a water pump, the method comprising: obtaining audio data when the water pump is working, the audio data including respective audio data of the water pump and the water in the water pump, wherein the audio data of the water pump is first audio data and the audio data of the water is second audio data; extracting the first audio data from the audio data by using a trained voiceprint extraction model; comparing the first audio data with audio sample data stored in a voiceprint database to determine the operating state of the water pump; and when it is determined that the operating state of the water pump is a fault, determining the fault type of the fault.

[0006] In an embodiment, the method further comprises: after obtaining the audio data when the water pump is working, first identifying the respective audio data of the water pump and the water in the water pump in the audio data by using a trained voiceprint recognition model, and then extracting the first audio data from the identified audio data by using a trained voiceprint extraction model.

[0007] In one embodiment, identifying the respective audio data of the water pump and the water in the water pump in the audio data by using a trained voiceprint recognition model includes: extracting the spectral features of the audio data by using the trained voiceprint recognition model; and identifying the respective audio data of the water pump and the water in the water pump in the audio data based on the spectral features.

[0008] In one embodiment, the audio sample data includes positive sample data and negative sample data, where the positive sample data is the audio data when the water pump is operating normally, and the negative sample data is the audio data when the water pump fails. The method further includes: comparing the first audio data with the positive sample data to determine whether the first audio data is within the range of the positive sample data, so as to determine the operating state of the water pump.

[0009] In one embodiment, the method further includes: converting the positive sample data into a positive sample voiceprint spectrogram; converting the first audio data into a first voiceprint spectrogram; and comparing the first voiceprint spectrogram with the positive sample voiceprint spectrogram to determine the coincidence degree between the first voiceprint spectrogram and the positive sample voiceprint spectrogram, so as to determine the operating state of the water pump.

[0010] In one embodiment, the method further includes: determining the left risk value and the right risk value of the positive sample voiceprint spectrogram; determining the left risk value and the right risk value of the first voiceprint spectrogram; and comparing the left risk value and the right risk value of the first voiceprint spectrogram with the left risk value and the right risk value of the positive sample voiceprint spectrogram respectively to determine the operating state of the water pump.

[0011] In one embodiment, the method further includes: when determining that the operating state of the water pump is a failure, comparing the first audio data with the negative sample data to determine the type of the failure.

[0012] In one embodiment, the method further includes: converting the negative sample data into a negative sample voiceprint spectrogram; converting the first audio data into a first voiceprint spectrogram; and comparing the first voiceprint spectrogram with the negative sample voiceprint spectrogram to determine the coincidence degree between the first voiceprint spectrogram and the negative sample voiceprint spectrogram, so as to determine the type of the failure.

[0013] In one embodiment, the method further includes: determining the left deviation value and the right deviation value of the negative sample voiceprint spectrogram; determining the left deviation value and the right deviation value of the first voiceprint spectrogram; and comparing the left deviation value and the right deviation value of the first voiceprint spectrogram with the left deviation value and the right deviation value of each negative sample voiceprint spectrogram respectively to determine the type of the failure.

[0014] In one embodiment, the method further includes: when the first voiceprint spectrogram does not coincide with any of the negative sample voiceprint spectrograms, determining that the first audio data indicates a new fault type, and storing the first audio data as new negative sample data in the voiceprint database.

[0015] According to another embodiment of the present invention, there is provided a computing device, which includes a memory and a processor. A computer program is stored on the memory. When the computer program is run by the processor, the processor is caused to execute the method as described above.

[0016] According to still another embodiment of the present invention, there is provided a computer-readable medium, on which a computer program is stored. When the computer program is run, it executes the method as described above.

[0017] The method, computing device, and computer-readable medium for determining the operating state and fault type of a water pump according to an embodiment of the present invention can determine the operating state and fault type of the water pump by automatically collecting the audio during the operation of the water pump, performing noise reduction processing on the audio, and analyzing it, greatly improving the recognition efficiency, comprehensiveness, and inheritance of water pump faults and their types. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The following drawings of the present invention are hereby incorporated as part of the present invention for understanding the present invention. The embodiments of the present invention and their descriptions are shown in the drawings to explain the principles of the present invention.

[0019] In the drawings:

[0020] Figure 1 is a schematic structural block diagram of an electronic device for implementing the method and computing device for determining the operating state and fault type of a water pump according to an embodiment of the present invention;

[0021] Figure 2 is an exemplary step flowchart of the method for determining the operating state and fault type of a water pump according to an embodiment of the present invention.

[0022] Figure 3 shows a schematic structural block diagram of a computing device according to another embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0023] To make the objectives, technical solutions, and advantages of the present invention more apparent, exemplary embodiments according to the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments of the present invention. It should be understood that the present invention is not limited by the exemplary embodiments described herein. Based on the embodiments of the present invention described herein, all other embodiments obtained by those skilled in the art without creative efforts shall fall within the protection scope of the present invention.

[0024] As described above, in the existing method of discovering the location and type of pump failures through manual inspections, the discovery cycle is long, and it is prone to misinspections and omissions, and it relies on expert experience.

[0025] Therefore, in order to quickly and accurately discover the location and type of pump failures, the present invention provides a method for determining the operating state and failure type of a pump. The method includes: obtaining audio data when the pump is operating, where the audio data includes the respective audio data of the pump and the water in the pump, where the audio data of the pump is the first audio data, and the audio data of the water is the second audio data; using a trained voiceprint extraction model to extract the first audio data from the audio data; comparing the first audio data with the audio sample data stored in a voiceprint database to determine the operating state of the pump; and when determining that the operating state of the pump is a failure, determining the failure type of the failure.

[0026] According to the method for determining the operating state and failure type of a pump of the present invention, by automatically collecting the audio when the pump is operating, performing noise reduction processing on the audio and analyzing it, the operating state and failure type of the pump can be determined, greatly improving the recognition efficiency, comprehensiveness, and inheritance of pump failures and their types.

[0027] The following describes in detail the method and computing device for determining the operating state and failure type of a pump according to the present invention with reference to specific embodiments.

[0028] First, refer to Figure 1 to describe the electronic device 100 for implementing the method and computing device for determining the operating state and failure type of a pump according to the embodiments of the present invention.

[0029] In one embodiment, the electronic device 100 can be, for example, a laptop computer, a desktop computer, a tablet computer, a learning machine, a mobile device (such as a smart phone, a smart watch, etc.), an embedded computer, a tower server, a rack server, a blade server, or any other suitable electronic device.

[0030] In one embodiment, the electronic device 100 can include at least one processor 102 and at least one memory 104.

[0031] Among them, the memory 104 can be a volatile memory, such as a random access memory (RAM), a cache memory, a dynamic random access memory (DRAM) (including stacked DRAM), or a high bandwidth memory (HBM), etc., or it can be a non-volatile memory, such as a read-only memory (ROM), a flash memory, 3D Xpoint, etc. In one embodiment, some parts of the memory 104 can be volatile memory, while another part can be non-volatile memory (for example, using a two-level memory hierarchy). The memory 104 is used to store a computer program, which, when run, can implement the client functions in the embodiments of the present invention described below (implemented by the processor) and / or other desired functions.

[0032] The processor 102 can be a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor, or other processing units with data processing capabilities and / or instruction execution capabilities. The processor 102 can be communicatively coupled via a communication bus to any suitable number or type of components, peripherals, modules, or devices. In one embodiment, the communication bus can be implemented using any suitable protocol, such as Peripheral Component Interconnect (PCI), Peripheral Component Interconnect Express (PCIe), Accelerated Graphics Port (AGP), HyperTransport, or any other bus or one or more point-to-point communication protocols.

[0033] The electronic device 100 may further include an input device 106 and an output device 108. Among them, the input device 106 is a device for receiving user input, which may include a keyboard, a mouse, a touchpad, a microphone, etc. In addition, the input device 106 can also be any interface for receiving information. The output device 108 can output various information (such as images or sounds) to the outside (such as the user), and it can include one or more of a display, a speaker, etc.

[0034] Next, refer to Figure 2 An exemplary step flowchart of a method 200 for determining the operating state and fault type of a water pump according to an embodiment of the present invention is described.

[0035] As Figure 2 shown, the method 200 for determining the operating state and fault type of a water pump may include the following steps:

[0036] In step S210, audio data when the water pump is working is acquired, and the audio data includes the respective audio data of the water pump and the water in the water pump, where the audio data of the water pump is the first audio data, and the audio data of the water is the second audio data.

[0037] In step S220, the first audio data is extracted from the audio data by using a trained voiceprint extraction model.

[0038] In step S230, the first audio data is compared with the audio sample data stored in the voiceprint database to determine the operating state of the water pump.

[0039] In step S240, when it is determined that the operating state of the water pump is a fault, the type of the fault is determined.

[0040] In one embodiment, the audio data during the operation of the water pump can be collected in real time by using a sound probe deployed around the water pump. In one embodiment, the water pump can include a single-unit water pump, a double-unit water pump, etc. Exemplarily, the sound probe can include any type of sound sensor well-known in the art, such as a piezoelectric ceramic sensor, a capacitive sensor, a magnetoelectric sensor, etc., and the present invention is not limited thereto. Exemplarily, the sound probe can be a microphone. Among them, the structure of the sound probe is well-known in the art and will not be described in detail herein.

[0041] In one embodiment, method 200 may further include: after obtaining the audio data during the operation of the water pump, first using a trained voiceprint recognition model to recognize the respective audio data of the water pump and the water in the water pump in the audio data, and then using a trained voiceprint extraction model to extract the first audio data from the recognized audio data.

[0042] In one embodiment, any suitable machine learning model well-known in the art can be used to construct a voiceprint recognition model, such as a text-independent type voiceprint modeling method, such as a GMM-UBM (Gaussian mixture model-universal background model), JFA (joint factor analysis) model, GMM-UBM i-vector model, supervised UBM i-vector model, deep neural network i-vector model, etc., and the present invention is not limited thereto.

[0043] Taking the construction of a voiceprint recognition model with a GMM-UBM model as an example, after constructing the voiceprint recognition model, a large number of target sounds and test sounds are collected, and spectral features (such as MFCC (Mel frequency cepstral coefficients)) are extracted through the voiceprint recognition model, and after repeated training and adaptive processing of a large amount of sound data (for example, using algorithms such as MAP (maximum a posteriori probability) algorithm, MLLR (maximum likelihood linear regression) algorithm, etc.) and confirmation decision, a voiceprint recognition model with a relatively high voiceprint recognition rate is obtained.

[0044] Then, the voiceprint recognition model is continuously trained with the sound of the water pump and the sound of the water, and the model parameters of the voiceprint recognition model are fine-tuned, and finally a trained voiceprint recognition model that can distinguish the sound of the water pump and the water therein is obtained.

[0045] In one embodiment, identifying the respective audio data of the water pump and the water in the water pump from the audio data by using a trained voiceprint recognition model may include: extracting spectral features of the audio data by using the trained voiceprint recognition model; and identifying the respective audio data of the water pump and the water in the water pump from the extracted spectral features.

[0046] In one embodiment, the extracted spectral features may be any suitable spectral features well known in the art, such as MFCC (Mel Frequency Cepstral Coefficient) features, Fbank (filter bank) features, BNF (Backus Normal Form) features, deep features, etc., and the present invention does not limit this.

[0047] Since the water pump and the water operate in the same environment simultaneously, the sound generated by the operation of the water pump is necessarily mixed with the sound of the water. Therefore, before analyzing and judging the acquired audio data, it is necessary to filter out the identified audio data of the water, that is, the second audio data, to remove the influence of the sound of the water and obtain the audio data of the water pump, that is, the first audio data. Embodiments of the present invention may use a trained voiceprint extraction model to extract the first audio data from the identified audio data.

[0048] In one embodiment, the following method may be used to train the voiceprint extraction model: converting the voiceprint turning point detection during the water filling of the water pump into a sequence labeling task for supervised model training. Among them, in the training data labeling stage, the audio frames at the turning points where the voiceprint turns are labeled with the label "1", while the audio frames at the turning points where the voiceprint does not turn are labeled with the label "0". Using the training data labeled by this method to train the voiceprint extraction model, so that the trained voiceprint extraction model can correctly obtain the positions of the audio frames where the turning points exist, and thus can correctly extract the audio data of the water pump.

[0049] In one embodiment, the voiceprint extraction model may be constructed by using a variety of neural networks, such as time delay neural network (TDNN), recurrent neural network (RNN), convolutional neural network (CNN), etc., and the present invention does not limit this.

[0050] In one embodiment, the voiceprint extraction model may be jointly trained with both the sound of the water pump and the sound of the water to improve the extraction effect of the voiceprint extraction model.

[0051] In one embodiment, the audio sample data may include positive sample data and negative sample data, where the positive sample data is the audio data when the water pump is working normally, and the negative sample data is the audio data when the water pump fails.

[0052] In one embodiment, method 200 may further include the step of pre - establishing a voiceprint database. In one embodiment, the voiceprint database may include a standard voiceprint database and a fault voiceprint database. Among them, the standard voiceprint database is used to store positive sample data when the water pump is operating normally, and the fault voiceprint database is used to store various types of negative sample data when various faults occur in the water pump.

[0053] In one embodiment, establishing the voiceprint database may include the following steps:

[0054] In step a, collect the sound when the water pump is operating normally and label it as positive sample data, and collect the sound when various faults occur in the water pump and label it as negative sample data;

[0055] In step b, store the positive sample data collected when the water pump is operating normally into the standard voiceprint database, and store the negative sample data collected when various types of faults occur in the water pump into the fault voiceprint database.

[0056] In one embodiment, various types of faults may include impeller damage, bearing wear, etc., and the present invention does not limit this. Among them, the negative sample data in the fault voiceprint database can be marked as corresponding negative sample data according to different fault types and stored separately.

[0057] In one embodiment, any well - known classification core algorithm in the art can be used to classify the fault types, such as Few - shot Learning, Active Learning, Transfer Learning, etc., and the present invention does not limit this.

[0058] In one embodiment, the positive sample data can be converted into a positive sample voiceprint spectrogram, the negative sample data in the fault voiceprint database can be converted into a negative sample voiceprint spectrogram, and the positive sample voiceprint spectrogram and the negative sample voiceprint spectrogram are stored in the standard voiceprint database.

[0059] In one embodiment, it is also possible to analyze and determine the left risk value and the right risk value of the positive sample voiceprint spectrogram and the left deviation value and the right deviation value of the negative sample voiceprint spectrogram, and store the left risk value and the right risk value of the positive sample voiceprint spectrogram and the left deviation value and the right deviation value of the negative sample voiceprint spectrogram in the standard voiceprint library.

[0060] In one embodiment, the left risk value and the right risk value of the positive sample voiceprint spectrogram can be determined in combination with the model parameters of the water pump, etc. They are respectively equal to the left boundary value and the right boundary value of the positive sample voiceprint spectrogram, and can also be slightly greater than or slightly less than the left boundary value and the right boundary value of the positive sample voiceprint spectrogram respectively.

[0061] In one embodiment, the left deviation value and the right deviation value of the negative sample voiceprint spectrogram can be respectively equal to the left boundary value and the right boundary value of the negative sample voiceprint spectrogram, or can be respectively slightly greater than or slightly less than the left boundary value and the right boundary value of the negative sample voiceprint spectrogram.

[0062] In one embodiment, comparing the first audio data for the water pump with the audio sample data stored in the voiceprint database may include: comparing the first audio data with the positive sample data in the standard voiceprint database to determine whether the first audio data is within the range of the positive sample data, so as to determine the operating state of the water pump. Among them, when the first audio data is within the range of the positive sample data, it is determined that the operating state of the water pump is normal; when the first audio data is not within the range of the positive sample data, it is determined that the operating state of the water pump is faulty.

[0063] In one embodiment, comparing the first audio data for the water pump with the audio sample data stored in the voiceprint database may further include: converting the first audio data into a first voiceprint spectrogram, and comparing the first voiceprint spectrogram with the positive sample voiceprint spectrogram to determine the coincidence degree between the first voiceprint spectrogram and the positive sample voiceprint spectrogram, so as to determine the operating state of the water pump. Among them, when the coincidence degree between the first voiceprint spectrogram and the positive sample voiceprint spectrogram is greater than or equal to a preset coincidence degree threshold, it is determined that the operating state of the water pump is normal; when the coincidence degree between the first voiceprint spectrogram and the positive sample voiceprint spectrogram is less than the preset coincidence degree threshold, it is determined that the operating state of the water pump is faulty.

[0064] In one embodiment, comparing the first audio data for the water pump with the audio sample data stored in the voiceprint database may further include: determining the left risk value and the right risk value of the first voiceprint spectrogram; comparing the left risk value and the right risk value of the first voiceprint spectrogram with the left risk value and the right risk value of the positive sample voiceprint spectrogram respectively to determine the operating state of the water pump. Among them, the left risk value and the right risk value of the first voiceprint spectrogram can be respectively equal to the left boundary value and the right boundary value of the first voiceprint spectrogram, or can be respectively slightly greater than or slightly less than the left boundary value and the right boundary value of the first voiceprint spectrogram. Among them, when the left risk value of the first voiceprint spectrogram is greater than or equal to the left risk value of the positive sample voiceprint spectrogram and / or the right risk value of the first voiceprint spectrogram is less than or equal to the right risk value of the positive sample voiceprint spectrogram, it is determined that the operating state of the water pump is normal; when both the left risk value and the right risk value of the first voiceprint spectrogram are less than the left risk value of the positive sample voiceprint spectrogram, or both the left risk value and the right risk value of the first voiceprint spectrogram are greater than the right risk value of the positive sample voiceprint spectrogram, it is determined that the operating state of the water pump is faulty.

[0065] In one embodiment, when determining that the operating state of the water pump is a fault, determining the type of the fault may include: when determining that the operating state of the water pump is a fault, comparing the first audio data with various types of negative sample data in the fault voiceprint database to determine the type of the fault. Wherein, when the first audio data is within the range of a certain type of negative sample data in the fault voiceprint database, it is determined that the water pump has the type of fault corresponding to the negative sample data.

[0066] In one embodiment, when determining that the operating state of the water pump is a fault, determining the type of the fault may further include: converting the first audio data into a first voiceprint spectrogram, comparing the first voiceprint spectrogram with the negative sample voiceprint spectrogram to determine the coincidence degree between the first voiceprint spectrogram and the negative sample voiceprint spectrogram, so as to determine the type of the water pump fault. Wherein, when the coincidence degree between the first voiceprint spectrogram and a certain negative sample voiceprint spectrogram is greater than or equal to a preset coincidence degree threshold, it is determined that the water pump has the type of fault corresponding to the negative sample voiceprint spectrogram.

[0067] In one embodiment, when determining that the operating state of the water pump is a fault, determining the type of the fault may further include: determining the left deviation value and the right deviation value of the first voiceprint spectrogram, comparing the left deviation value and the right deviation value of the first voiceprint spectrogram with the left deviation value and the right deviation value of each negative sample voiceprint spectrogram respectively to determine the type of the fault. Wherein, the left risk value and the right risk value of the first voiceprint spectrogram may be respectively equal to the left boundary value and the right boundary value of the first voiceprint spectrogram, or may be slightly greater than or slightly less than the left boundary value and the right boundary value of the first voiceprint spectrogram respectively. Wherein, when the left deviation value of the first voiceprint spectrogram is greater than or equal to the left deviation value of a certain negative sample voiceprint spectrogram and the right deviation value of the first voiceprint spectrogram is less than or equal to the right deviation value of the negative sample voiceprint spectrogram, it is determined that the water pump has the type of fault corresponding to the negative sample voiceprint spectrogram.

[0068] In one embodiment, the following method may be adopted to analyze the positive sample data in the standard voiceprint database to determine the left risk value and the right risk value of the positive sample voiceprint spectrogram:

[0069] For the positive sample data in the standard voiceprint database, that is, the observation vector O, find the corresponding word sequence W such that the value of the conditional probability P(W|O) is the largest:

[0070] W = arg max(P(W|O))

[0071] According to the Bayes formula, it can be obtained:

[0072] P(W|O) = P(O|W)·P(W) / P(O)

[0073] Omitting P(O), we get:

[0074] W = arg max(P(O|W)·P(W))

[0075] Similarly, the above analysis method can also be used to analyze the negative sample data in the fault voiceprint database. According to the above analysis results, the left risk value and right risk value of the positive sample voiceprint spectrogram, as well as the left deviation value and right deviation value of the negative sample voiceprint spectrogram, can be determined.

[0076] In one embodiment, method 200 may further include: when the first voiceprint spectrogram does not coincide with any of the negative sample voiceprint spectrograms, determining that the first audio data indicates a new fault type, and storing the first audio data as new negative sample data in the voiceprint database, specifically, storing it in the fault voiceprint database.

[0077] According to the method for determining the operating state and fault type of a water pump according to the present invention, by automatically collecting the audio during the operation of the water pump, performing noise reduction processing on the audio and analyzing it, the operating state and fault type of the water pump can be determined, greatly improving the recognition efficiency, comprehensiveness, and inheritance of water pump faults and their types.

[0078] In yet another embodiment, the present invention provides a computing device. Referring to Figure 3 , Figure 3 shows a schematic structural block diagram of a computing device 300 according to another embodiment of the present invention. As Figure 3 shown, the computing device 300 may include a memory 310 and a processor 320, where a computer program is stored on the memory 310, and when the computer program is run by the processor 320, the processor 320 is caused to execute the method 200 for determining the operating state and fault type of a water pump as described above.

[0079] Those skilled in the art can understand the specific operations of the computing device 300 according to the embodiments of the present invention in combination with the content described above. For the sake of brevity, specific details are not described here again, and only some main operations of the processor 320 are described as follows:

[0080] Obtain audio data during the operation of the water pump, where the audio data includes the respective audio data of the water pump and the water in the water pump, where the audio data of the water pump is the first audio data and the audio data of the water is the second audio data;

[0081] Extract the first audio data from the audio data by using a trained voiceprint extraction model;

[0082] Compare the first audio data with the audio sample data stored in the voiceprint database to determine the operating state of the water pump; and

[0083] When it is determined that the operating state of the water pump is a fault, determine the type of the fault.

[0084] According to the computing device of an embodiment of the present invention, by automatically collecting the audio of the water pump during operation, performing noise reduction processing on the audio and analyzing it, the operating state and fault type of the water pump can be determined, greatly improving the recognition efficiency, comprehensiveness and inheritance of water pump faults and their types.

[0085] In another embodiment, the present invention provides a computer-readable medium on which a computer program is stored, and the computer program executes the method 200 for determining the operating state and fault type of the water pump as described in the above embodiment when running. Any tangible, non-transitory computer-readable medium can be used, including magnetic storage devices (hard disks, floppy disks, etc.), optical storage devices (CD-ROMs, DVDs, Blu-ray discs, etc.), flash memories and / or the like. These computer program instructions can be loaded onto a general-purpose computer, a special-purpose computer or other programmable data processing devices to form a machine, so that the instructions executed on the computer or other programmable data processing devices can generate a device for realizing the specified function. These computer program instructions can also be stored in a computer-readable memory, and the computer-readable memory can instruct the computer or other programmable data processing devices to operate in a specific manner, so that the instructions stored in the computer-readable memory can form a manufactured article including a device for realizing the specified function. The computer program instructions can also be loaded onto the computer or other programmable data processing devices, so as to execute a series of operation steps on the computer or other programmable devices to generate a computer-implemented process, so that the instructions executed on the computer or other programmable devices can provide steps for realizing the specified function.

[0086] Although example embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above example embodiments are merely exemplary and are not intended to limit the scope of the present invention thereto. Those of ordinary skill in the art can make various changes and modifications therein without departing from the scope and spirit of the present invention. All such changes and modifications are intended to be included within the scope of the present invention as claimed in the appended claims.

[0087] In the specification provided herein, a large number of specific details are set forth. However, it can be understood that the embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0088] Similarly, it should be understood that, for the purpose of streamlining the present invention and facilitating the understanding of one or more of the various inventive aspects, in the description of the exemplary embodiments of the present invention, the various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, the method of the present invention should not be construed as reflecting an intention that the claimed invention requires more features than those expressly recited in each claim. Rather, as reflected by the corresponding claims, the inventive point lies in that the corresponding technical problem can be solved by features less than all the features of a single disclosed embodiment. Therefore, the claims following the detailed description are hereby expressly incorporated into the detailed description, where each claim itself serves as a separate embodiment of the present invention.

[0089] Those skilled in the art will appreciate that, except where features are mutually exclusive, any combination can be employed of all the features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all the processes or units of any method or apparatus so disclosed. Each feature disclosed in this specification (including the accompanying claims, abstract, and drawings), unless otherwise expressly stated, may be replaced by alternative features serving the same, equivalent, or similar purpose.

[0090] Furthermore, those skilled in the art will be able to understand that, although some of the embodiments described herein include certain features included in other embodiments but not other features, combinations of the features of different embodiments are meant to be within the scope of the present invention and form different embodiments. For example, in the claims, any one of the claimed embodiments can be used in any combination.

[0091] It should be noted that the above embodiments illustrate rather than limit the present invention, and those skilled in the art can design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention can be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In a unit claim listing several devices, several of these devices can be embodied by the same item of hardware. The use of the words first, second, and third, etc. does not denote any order. These words can be interpreted as names.

[0092] As described above, it is only the specific implementation manner of the present invention or the description of the specific implementation manner. The protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. The protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for determining the operating state and fault type of a water pump, characterized in that The method includes: Obtaining audio data when the water pump is working, where the audio data includes respective audio data of the water pump and the water in the water pump. Among them, the audio data of the water pump is the first audio data, and the audio data of the water is the second audio data; Using a trained voiceprint extraction model to extract the first audio data from the audio data; Comparing the first audio data with the audio sample data stored in the voiceprint database to determine the operating state of the water pump; and When determining that the operating state of the water pump is a fault, determining the fault type of the fault; Where the audio sample data includes positive sample data, and the positive sample data is the audio data when the water pump is working properly. The comparing the first audio data with the audio sample data stored in the voiceprint database to determine the operating state of the water pump includes: Converting the positive sample data into a positive sample voiceprint spectrogram, and determining the left risk value and the right risk value of the positive sample voiceprint spectrogram; Converting the first audio data into a first voiceprint spectrogram, and determining the left risk value and the right risk value of the first voiceprint spectrogram; Comparing the left risk value and the right risk value of the first voiceprint spectrogram with the left risk value and the right risk value of the positive sample voiceprint spectrogram respectively to determine the operating state of the water pump.

2. The method according to claim 1, wherein The method further includes: after obtaining the audio data when the water pump is working, first using a trained voiceprint recognition model to identify the respective audio data of the water pump and the water in the water pump in the audio data, and then using a trained voiceprint extraction model to extract the first audio data from the identified audio data.

3. The method according to claim 2, wherein Among them, using a trained voiceprint recognition model to identify the respective audio data of the water pump and the water in the water pump in the audio data includes: Using a trained voiceprint recognition model to extract the spectral features of the audio data; and Based on the spectral features, identifying the respective audio data of the water pump and the water in the water pump in the audio data.

4. The method according to claim 1, wherein The method further includes: comparing the first audio data with the positive sample data to determine whether the first audio data is within the range of the positive sample data, so as to determine the operating state of the water pump.

5. The method according to claim 4, wherein The method further includes: Comparing the first voiceprint spectrogram with the positive sample voiceprint spectrogram to determine the coincidence degree between the first voiceprint spectrogram and the positive sample voiceprint spectrogram, so as to determine the operating state of the water pump.

6. The method according to claim 1, wherein The audio sample data further includes negative sample data, and the negative sample data is the audio data when the water pump fails. The method further includes: When determining that the operating state of the water pump is a fault, comparing the first audio data with the negative sample data to determine the fault type of the fault.

7. The method according to claim 6, wherein The method further includes: Converting the negative sample data into a negative sample voiceprint spectrogram; Converting the first audio data into a first voiceprint spectrogram; and Comparing the first voiceprint spectrogram with the negative sample voiceprint spectrogram to determine the coincidence degree between the first voiceprint spectrogram and the negative sample voiceprint spectrogram, so as to determine the fault type of the fault.

8. The method according to claim 7, characterized in that, The method further includes: Determine the left deviation value and the right deviation value of the negative sample voiceprint spectrogram; Determine the left deviation value and the right deviation value of the first voiceprint spectrogram; and Compare the left deviation value and the right deviation value of the first voiceprint spectrogram with the left deviation value and the right deviation value of each negative sample voiceprint spectrogram respectively to determine the fault type of the fault.

9. The method according to claim 7, wherein The method further includes: when the first voiceprint spectrogram does not coincide with any of the negative sample voiceprint spectrograms, determining that the first audio data indicates a new fault type, and storing the first audio data as new negative sample data in the voiceprint database.

10. A computing device, characterized in that, The computing device includes a memory and a processor, and a computer program is stored on the memory. When the computer program is run by the processor, the processor is caused to execute the method according to any one of claims 1-9.

11. A computer-readable medium, characterized in that, A computer program is stored on the computer-readable medium. When the computer program is run, it executes the method according to any one of claims 1-9.

Images