Deep learning-based precursor data fault diagnosis method and system, electronic equipment and storage medium

Abstract

The invention provides a precursor data fault diagnosis method and system based on deep learning, electronic equipment and a storage medium. The method comprises the steps that on the basis of traditional time domain spectral analysis, precursor data of sensor vibration is added for primary processing, fault classification is carried out on specific faults of equipment through a VGG-11 convolutional neural network training model, manual spectrogram analysis is replaced, and corresponding reason analysis and solving measures are matched, so that fault diagnosis of the equipment is achieved. According to the scheme provided by the invention, the vibration high alarm, the rotor imbalance fault, the gear wear fault and the bearing misalignment fault of the equipment can be classified.

Description

Technical Field

[0001] The present invention belongs to the technical fields of fault diagnosis and deep learning, and particularly relates to a fault diagnosis method, system, electronic device and storage medium for precursor data based on deep learning. Background Art

[0002] The process of using various inspection and test methods to detect whether there are faults in a system and equipment is fault detection; and the process of further determining the location of the fault is fault location. The process of requiring the fault to be located at the product level that can be replaced during repair is called fault isolation. Fault diagnosis refers to the process of fault diagnosis and fault isolation.

[0003] Mechanical fault diagnosis studies the reflection of changes in the operating state of a machine or unit in diagnostic information. The diagnostic objects include bearings, gears, rotors, and reciprocating machinery. It includes signal acquisition and sensing technology, the relationship between fault mechanisms and symptoms, signal processing and feature extraction, identification and classification, and intelligent decision-making, etc.

[0004] Time-domain and frequency-spectrum analysis is a commonly used signal processing technology, which can analyze vibration signals and conduct fault diagnosis. Frequency-spectrum analysis can obtain the frequency spectrum of vibration, and the frequency spectrum can be represented by a graph, with the horizontal axis indicating frequency and the vertical axis indicating amplitude. Fault diagnosis can be carried out by manually analyzing the frequency spectrum.

[0005] In the field of fault diagnosis, the acquisition, processing, and analysis of a large amount of data are crucial. With the wide application of Internet technology and the popularization of intelligent devices, data sources such as sensors and monitoring devices have shown an explosive growth, which undoubtedly poses higher requirements for data acquisition and analysis. Compared with traditional fault diagnosis methods, deep learning technology can help people process a large amount of data more quickly and accurately, so as to achieve targeted fault diagnosis, learn from a large number of fault cases, and improve the diagnostic accuracy.

[0006] Therefore, how to provide a fault diagnosis method, system, electronic device and storage medium for precursor data based on deep learning has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0007] The purpose of the present invention is to provide a fault diagnosis method, system, electronic device and storage medium for precursor data based on deep learning.

[0008] According to the first aspect of the present invention, there is provided a fault diagnosis for precursor data based on deep learning, including,

[0009] Step S1: Determine whether it is a vibration high-alarm fault according to the average value of the vibration signal collected by the sensor in a certain second; if it is a vibration high-alarm fault, the algorithm stops, otherwise go to step S2;

[0010] Step S2: Determine whether it is a gear wear fault trend based on the effective value, peak-to-peak value, amplitude of the first harmonic of the meshing frequency, amplitude of the second harmonic of the meshing frequency, and amplitude of the third harmonic of the meshing frequency of the vibration signal; if it is a gear wear fault trend, go to step S5; otherwise, go to step S3;

[0011] Step S3: Determine whether it is a rotor imbalance fault trend based on the effective value, peak value, kurtosis, amplitude of the first harmonic in the X direction, amplitude of the first harmonic in the Y direction, amplitude of the first harmonic of the vibration signal in the X or Y direction, amplitude of the second harmonic, and amplitude of the third harmonic of the vibration signals in the X and Y directions; if it is a rotor imbalance fault trend, go to step S5; otherwise, go to step S4;

[0012] Step S4: Determine whether it is a bearing misalignment fault trend based on the effective value, peak-to-peak value, and amplitude of the second rotational speed frequency of the vibration displacement of the shaft at +45 degrees or -45 degrees of the vibration signal; if it is a bearing misalignment fault trend, go to step S5;

[0013] Step S5: Obtain the fault type by using a deep learning network based on the VGG-11 network according to the vibration signals collected in a certain second and the previous second.

[0014] According to the method of the first aspect of the present invention, in the step S1, the determining whether it is a vibration high alarm fault according to the average value of the vibration signals collected by the sensor in a certain second includes:

[0015] If the average value of the vibration signals collected by the sensor in a certain second is greater than the upper limit of the vibration signals, it is determined as a vibration high alarm fault.

[0016] According to the method of the first aspect of the present invention, in the step S2, the determining whether it is a gear wear fault trend according to the effective value, peak-to-peak value, amplitude of the first harmonic of the meshing frequency, amplitude of the second harmonic of the meshing frequency, and amplitude of the third harmonic of the meshing frequency of the vibration signal includes:

[0017] If the effective value of the vibration signal > the upper limit of the vibration signal and the peak-to-peak value of the vibration signal > the upper limit of the vibration signal and the amplitude of the first harmonic of the meshing frequency, the amplitude of the second harmonic of the meshing frequency, and the amplitude of the third harmonic of the meshing frequency of the vibration signal > the upper limit of the meshing frequency amplitude are true, it is determined as a gear wear fault trend.

[0018] According to the method of the first aspect of the present invention, in the step S3, the determining whether it is a rotor imbalance fault trend according to the effective value, peak value, kurtosis, amplitude of the first harmonic in the X direction, amplitude of the first harmonic in the Y direction, amplitude of the first harmonic of the vibration signal in the X or Y direction, amplitude of the second harmonic, and amplitude of the third harmonic of the vibration signals in the X and Y directions includes:

[0019] The upper limit of the ratio of the effective value of the vibration signals in the X and Y directions to the peak value of the vibration signals in the X and Y directions, and the kurtosis is greater than the upper limit of kurtosis, and the amplitude of the first harmonic of the vibration signal in the X direction is greater than the amplitude of the first harmonic of the vibration signal in the Y direction, and the amplitude of the first harmonic of the vibration signal in the X or Y direction is greater than the amplitude of the second harmonic of the vibration signal in the X or Y direction, and the amplitude of the first harmonic of the vibration signal in the X or Y direction is greater than the amplitude of the third harmonic of the vibration signal in the X or Y direction, and the upper limit of the amplitude of the first harmonic of the vibration signal in the X or Y direction is true, then it is judged as a rotor imbalance fault trend.

[0020] According to the method of the first aspect of the present invention, in the step S4, the judging whether it is a bearing misalignment fault trend according to the effective value, peak-to-peak value and double-rotational speed frequency amplitude of the vibration displacement of the shaft +45 degrees or shaft -45 degrees of the vibration signal includes:

[0021] If the effective value of the vibration displacement of the shaft +45 degrees or shaft -45 degrees of the vibration signal is greater than the upper limit of the vibration displacement, and the peak-to-peak value of the vibration displacement is greater than the upper limit of the vibration displacement, and the double-rotational speed frequency amplitude of the vibration displacement is greater than the upper limit of the vibration displacement is true, then it is judged as a bearing misalignment fault trend.

[0022] According to the method of the first aspect of the present invention, in the step S4, the obtaining of the fault type by using a deep learning network based on the VGG-11 network according to the vibration signals collected in a certain second and the previous second includes:

[0023] According to the vibration signals collected in a certain second and the previous second, draw a time-domain curve;

[0024] Subtract the average value of the vibration signals collected in a certain second and the previous second from the vibration signals collected in a certain second and the previous second to obtain the transformed vibration signals;

[0025] Perform FFT transformation on the transformed vibration signals to obtain a frequency spectrum curve;

[0026] Input the time-domain curve and the frequency spectrum curve into a deep learning network based on the VGG-11 network for classification to obtain the fault type.

[0027] According to the method of the first aspect of the present invention, in the step S4, the deep learning network based on the VGG-11 network is;

[0028] Replace the max pooling in the VGG-11 network with average pooling.

[0029] The second aspect of the present invention discloses a premonitory data fault diagnosis system based on deep learning, and the system includes:

[0030] The first processing module is configured to determine whether it is a vibration high alarm fault according to the average value of the vibration signal collected by the sensor in a certain second; if it is a vibration high alarm fault, the system stops the determination, otherwise it goes to the second processing module;

[0031] The second processing module is configured to determine whether it is a gear wear fault trend according to the effective value, peak-to-peak value, amplitude of the first-order meshing frequency, amplitude of the second-order meshing frequency, and amplitude of the third-order meshing frequency of the vibration signal; if it is a gear wear fault trend, it goes to the fifth processing module; otherwise it goes to the third processing module;

[0032] The third processing module is configured to determine whether it is a rotor imbalance fault trend according to the effective value, peak value, kurtosis, amplitude of the first harmonic in the X direction, amplitude of the first harmonic in the Y direction, amplitude of the first harmonic of the vibration signal in the X or Y direction, amplitude of the second harmonic, and amplitude of the third harmonic of the vibration signals in the X and Y directions; if it is a rotor imbalance fault trend, it goes to the fifth processing module; otherwise it goes to the fourth processing module;

[0033] The fourth processing module is configured to determine whether it is a bearing misalignment fault trend according to the effective value, peak-to-peak value, and amplitude of the second rotational speed frequency of the vibration displacement of the shaft +45 degrees or shaft -45 degrees of the vibration signal; if it is a bearing misalignment fault trend, it goes to the fifth processing module;

[0034] The fifth processing module is configured to obtain the fault type by using a deep learning network based on the VGG-11 network according to the vibration signals collected in the certain second and the previous second.

[0035] A third aspect of the present invention discloses an electronic device. The electronic device includes a memory and a processor. The memory stores a computer program. When the processor executes the computer program, the steps in a method for diagnosing faults in precursor data based on deep learning according to any one of the first aspects of the present disclosure are implemented.

[0036] A fourth aspect of the present invention discloses a computer-readable storage medium. A computer program is stored on the computer-readable storage medium. When the computer program is executed by a processor, the steps in a method for diagnosing faults in precursor data based on deep learning according to any one of the first aspects of the present disclosure are implemented.

[0037] The beneficial effects brought by the present invention are as follows:

[0038] As can be seen from the above solution, the embodiments of the present invention provide a method, a system, an electronic device, and a storage medium for diagnosing faults in precursor data based on deep learning, and have the following beneficial effects: improving the accuracy of fault diagnosis and improving the automation level of fault diagnosis. Description of the Drawings

[0039] Figure 1 Flow chart of a premonitory data fault diagnosis method based on deep learning provided according to an embodiment;

[0040] Figure 2 Structural diagram of a premonitory data fault diagnosis system based on deep learning according to an embodiment of the present invention;

[0041] Figure 3 Structural diagram of an electronic device according to an embodiment of the present invention. Detailed implementation manners

[0042] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0043] Embodiment 1:

[0044] According to the first aspect of the present invention, the present invention discloses a premonitory data fault diagnosis method based on deep learning. Figure 1 As a flow chart of a premonitory data fault diagnosis method based on deep learning according to an embodiment of the present invention, as Figure 1 shown, the method includes:

[0045] Step S1: Determine whether it is a vibration high alarm fault according to the average value of the vibration signal collected by the sensor in a certain second; if it is a vibration high alarm fault, the algorithm stops, otherwise go to step S2;

[0046] Step S2: Determine whether it is a gear wear fault trend according to the effective value, peak-to-peak value, amplitude of the meshing frequency by one time, amplitude of the meshing frequency by two times, and amplitude of the meshing frequency by three times of the vibration signal; if it is a gear wear fault trend, go to step S5; otherwise go to step S3;

[0047] Step S3: Determine whether it is a rotor imbalance fault trend according to the effective value, peak value, kurtosis, amplitude of the first harmonic in the X direction, amplitude of the first harmonic in the Y direction, amplitude of the first harmonic of the vibration signal in the X or Y direction, amplitude of the second harmonic, and amplitude of the third harmonic of the vibration signals in the X and Y directions; if it is a rotor imbalance fault trend, go to step S5; otherwise go to step S4;

[0048] Step S4. Determine whether it is a bearing misalignment fault trend based on the effective value, peak-to-peak value, and twice the rotational speed frequency amplitude of the vibration displacement at the +45-degree or -45-degree axis of the vibration signal; if it is a bearing misalignment fault trend, go to step S5;

[0049] Step S5. Based on the vibration signals collected in a certain second and the previous second, use a deep learning network based on the VGG-11 network to obtain the fault type.

[0050] In step S1, determine whether it is a vibration high alarm fault based on the average value of the vibration signal collected by the sensor in a certain second; if it is a vibration high alarm fault, the algorithm stops, otherwise go to step S2.

[0051] In some embodiments, in step S1, the determination of whether it is a vibration high alarm fault based on the average value of the vibration signal collected by the sensor in a certain second includes:

[0052] Perform preprocessing on the vibration data measurement points of the sensor. If the average value of the vibration signal collected by the sensor in a certain second is greater than the upper limit of the vibration signal, it is determined as a vibration high alarm fault.

[0053] In step S2, determine whether it is a gear wear fault trend based on the effective value, peak-to-peak value, first harmonic frequency amplitude, second harmonic frequency amplitude, and third harmonic frequency amplitude of the vibration signal; if it is a gear wear fault trend, go to step S5; otherwise go to step S3.

[0054] In some embodiments, in step S2, the determination of whether it is a gear wear fault trend based on the effective value, peak-to-peak value, first harmonic frequency amplitude, second harmonic frequency amplitude, and third harmonic frequency amplitude of the vibration signal includes:

[0055] If the effective value of the vibration signal > the upper limit of the vibration signal and the peak-to-peak value of the vibration signal > the upper limit of the vibration signal and the first harmonic frequency amplitude, second harmonic frequency amplitude, and third harmonic frequency amplitude of the vibration signal > the upper limit of the harmonic frequency amplitude are true, it is determined as a gear wear fault trend.

[0056] In step S3, determine whether it is a rotor imbalance fault trend based on the effective value, peak value, kurtosis, first harmonic frequency amplitude in the X direction, first harmonic frequency amplitude in the Y direction, first harmonic frequency amplitude of the vibration signal in the X or Y direction, second harmonic frequency amplitude, and third harmonic frequency amplitude of the vibration signals in the X and Y directions; if it is a rotor imbalance fault trend, go to step S5; otherwise go to step S4.

[0057] In some embodiments, in the step S3, determining whether there is a rotor imbalance fault trend based on the effective values, peak values, kurtosis, fundamental frequency amplitudes in the X direction, fundamental frequency amplitudes in the Y direction, fundamental frequency amplitudes of the vibration signal in the X or Y direction, second harmonic amplitudes, and third harmonic amplitudes of the vibration signals in the X and Y directions includes:

[0058] If the ratio of the effective value of the vibration signals in the X and Y directions to the peak value of the vibration signals in the X and Y directions > the upper limit of the ratio and the kurtosis > the upper limit of the kurtosis and the fundamental frequency amplitude of the vibration signal in the X direction > the fundamental frequency amplitude of the vibration signal in the Y direction and the fundamental frequency amplitude of the vibration signal in the X or Y direction > the second harmonic amplitude of the vibration signal in the X or Y direction and the fundamental frequency amplitude of the vibration signal in the X or Y direction > the third harmonic amplitude of the vibration signal in the X or Y direction and the fundamental frequency amplitude of the vibration signal in the X or Y direction > the upper limit of the frequency amplitude is true, then it is determined that there is a rotor imbalance fault trend.

[0059] In step S4, based on the effective value, peak-to-peak value, and double-rotational speed frequency amplitude of the vibration displacement at the shaft +45 degrees or shaft -45 degrees of the vibration signal, determine whether there is a bearing misalignment fault trend; if there is a bearing misalignment fault trend, then go to step S5.

[0060] In some embodiments, in the step S4, determining whether there is a bearing misalignment fault trend based on the effective value, peak-to-peak value, and double-rotational speed frequency amplitude of the vibration displacement at the shaft +45 degrees or shaft -45 degrees of the vibration signal includes:

[0061] If the effective value of the vibration displacement at the shaft +45 degrees or shaft -45 degrees of the vibration signal > the upper limit of the vibration displacement and the peak-to-peak value of the vibration displacement > the upper limit of the vibration displacement and the double-rotational speed frequency amplitude of the vibration displacement > the upper limit of the vibration displacement is true, then it is determined that there is a bearing misalignment fault trend.

[0062] In step S5, based on the vibration signals collected at a certain second and the previous second, use a deep learning network based on the VGG-11 network to obtain the fault type.

[0063] In some embodiments, in the step S5, obtaining the fault type by using a deep learning network based on the VGG-11 network according to the vibration signals collected at a certain second and the previous second includes:

[0064] Based on the vibration signals collected at a certain second and the previous second, draw a time-domain curve;

[0065] Subtract the average value of the vibration signals collected at a certain second and the previous second from the vibration signals collected at a certain second and the previous second to obtain the transformed vibration signals;

[0066] Perform FFT transformation on the transformed vibration signal to obtain a frequency spectrum curve;

[0067] Input the time-domain curve and the frequency spectrum curve into a deep learning network based on the VGG-11 network for classification to obtain the fault type.

[0068] The deep learning network based on the VGG-11 network is;

[0069] Replace the max pooling in the VGG-11 network with average pooling.

[0070] Specifically, the VGG-11 network consists of 8 convolutional layers and 3 fully connected layers, and a pooling layer follows each convolutional layer.

[0071] In some embodiments, match the phenomenon causes and treatment measures of the fault according to the diagnosed fault type.

[0072] In summary, the solution proposed by the present invention can classify high vibration alarms, rotor imbalance faults, gear wear faults, and bearing misalignment faults of equipment, and the classification accuracy rate is 92%.

[0073] Embodiment 2:

[0074] The present invention discloses a premonitory data fault diagnosis system based on deep learning. Figure 2 For a structural diagram of a premonitory data fault diagnosis system based on an embodiment of the present invention; as Figure 2 shown, the system 100 includes:

[0075] A first processing module 101, configured to judge whether it is a high vibration alarm fault according to the average value of the vibration signal collected by the sensor in a certain second; if it is a high vibration alarm fault, the system stops judging, otherwise it goes to the second processing module;

[0076] A second processing module 102, configured to judge whether it is a gear wear fault trend according to the effective value, peak-to-peak value, amplitude of the first meshing frequency, amplitude of the second meshing frequency, and amplitude of the third meshing frequency of the vibration signal; if it is a gear wear fault trend, it goes to the fifth processing module; otherwise it goes to the third processing module;

[0077] A third processing module 103, configured to judge whether it is a rotor imbalance fault trend according to the effective value, peak value, kurtosis, amplitude of the first harmonic in the X direction, amplitude of the first harmonic in the Y direction, amplitude of the first harmonic of the vibration signal in the X or Y direction, amplitude of the second harmonic, and amplitude of the third harmonic of the vibration signal in the X and Y directions; if it is a rotor imbalance fault trend, it goes to the fifth processing module; otherwise it goes to the fourth processing module;

[0078] The fourth processing module 104 is configured to determine whether there is a bearing misalignment fault trend based on the effective value, peak-to-peak value, and double rotational speed frequency amplitude of the vibration displacement at the +45-degree or -45-degree axis of the vibration signal; if there is a bearing misalignment fault trend, it proceeds to the fifth processing module;

[0079] The fifth processing module 105 is configured to obtain the fault type by using a deep learning network based on the VGG-11 network according to the vibration signals collected in a certain second and the previous second.

[0080] For the system according to the second aspect of the present invention, the first processing module 101 is specifically configured to determine whether it is a vibration high alarm fault according to the average value of the vibration signal collected by the sensor in a certain second, including:

[0081] Perform preprocessing on the vibration data measurement point of the sensor. If the average value of the vibration signal collected by the sensor in a certain second is greater than the upper limit of the vibration signal, it is determined as a vibration high alarm fault.

[0082] For the system according to the second aspect of the present invention, the second processing module 102 is specifically configured to determine whether there is a gear wear fault trend according to the effective value, peak-to-peak value, single meshing frequency amplitude, double meshing frequency amplitude, and triple meshing frequency amplitude of the vibration signal, including:

[0083] If the effective value of the vibration signal > the upper limit of the vibration signal and the peak-to-peak value of the vibration signal > the upper limit of the vibration signal and the single meshing frequency amplitude, double meshing frequency amplitude, and triple meshing frequency amplitude of the vibration signal > the upper limit of the meshing frequency amplitude are true, it is determined as a gear wear fault trend.

[0084] For the system according to the second aspect of the present invention, the third processing module 103 is specifically configured to determine whether there is a rotor imbalance fault trend according to the effective value, peak value, kurtosis, single-frequency amplitude in the X direction, single-frequency amplitude in the Y direction, single-frequency amplitude of the vibration signal in the X or Y direction, double-frequency amplitude, and triple-frequency amplitude of the vibration signals in the X and Y directions, including:

[0085] The effective value of the vibration signals in the X and Y directions / the peak value of the vibration signals in the X and Y directions > the ratio upper limit and the kurtosis > the kurtosis upper limit and the single-frequency amplitude of the vibration signal in the X direction > the single-frequency amplitude of the vibration signal in the Y direction and the single-frequency amplitude of the vibration signal in the X or Y direction > the double-frequency amplitude of the vibration signal in the X or Y direction and the single-frequency amplitude of the vibration signal in the X or Y direction > the triple-frequency amplitude of the vibration signal in the X or Y direction and the single-frequency amplitude of the vibration signal in the X or Y direction > the frequency amplitude upper limit are true, it is determined as a rotor imbalance fault trend.

[0086] The system according to the second aspect of the present invention, the fourth processing module 104 is specifically configured to determine whether it is a bearing misalignment fault trend according to the effective value, peak-to-peak value, and double rotational speed frequency amplitude of the vibration displacement of the shaft +45 degrees or shaft -45 degrees of the vibration signal, including:

[0087] If the effective value of the vibration displacement of the shaft +45 degrees or shaft -45 degrees of the vibration signal > the upper limit of the vibration displacement and the peak-to-peak value of the vibration displacement > the upper limit of the vibration displacement and the double rotational speed frequency amplitude of the vibration displacement > the upper limit of the vibration displacement is true, then it is determined as a bearing misalignment fault trend.

[0088] The system according to the second aspect of the present invention, the fifth processing module 105 is specifically configured to obtain the fault type according to the vibration signals collected in a certain second and the previous second by using a deep learning network based on the VGG-11 network, including:

[0089] According to the vibration signals collected in a certain second and the previous second, draw a time-domain curve;

[0090] Subtract the average value of the vibration signals collected in a certain second and the previous second from the vibration signals collected in a certain second and the previous second to obtain the transformed vibration signals;

[0091] Perform FFT transformation on the transformed vibration signals to obtain a frequency spectrum curve;

[0092] Input the time-domain curve and the frequency spectrum curve into a deep learning network based on the VGG-11 network for classification to obtain the fault type.

[0093] The deep learning network based on the VGG-11 network is;

[0094] Replace the max pooling in the VGG-11 network with average pooling.

[0095] Specifically, the VGG-11 network consists of 8 convolutional layers and 3 fully connected layers, and a pooling layer follows each convolutional layer.

[0096] Embodiment 3:

[0097] This application discloses an electronic device. The electronic device includes a memory and a processor. When the processor executes the computer program stored in the memory, it implements the steps in any one of the deep learning-based precursor data fault diagnosis methods in the disclosed Embodiment 1 of the present invention.

[0098] Figure 3 For a structural diagram of an electronic device according to an embodiment of the present invention, as Figure 3As shown in the figure, the electronic device includes a processor, a memory, a communication interface, a display screen, and an input device connected through a system bus. Among them, the processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The communication interface of the electronic device is used to communicate with external terminals in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a carrier network, near field communication (NFC), or other technologies. The display screen of the electronic device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the electronic device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the electronic device, or an external keyboard, touchpad, or mouse, etc.

[0099] Those skilled in the art can understand that Figure 3 the structure shown in is only the structure diagram of the part related to the technical solution of the present disclosure, and does not constitute a limitation on the electronic device to which the solution of the present application is applied. The specific electronic device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0100] Embodiment 4:

[0101] The present invention discloses a computer-readable storage medium. A computer program is stored on the computer-readable storage medium. When the computer program is executed by a processor, the steps in a method for diagnosing precursor data faults based on deep learning according to any one of Embodiment 1 of the present invention are implemented.

[0102] Please note that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combinations of these technical features do not conflict, they should all be considered as within the scope described in this specification. The above embodiments only represent several implementation manners of the present application, and their descriptions are relatively specific and detailed, but they should not be construed as a limitation on the scope of the invention patent. It should be pointed out that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

[0103] The embodiments of the subject matter and the functional operations described in this specification can be implemented in the following: digital electronic circuits, tangible computer software or firmware, computer hardware including the structures disclosed in this specification and their structural equivalents, or a combination of one or more of them. Embodiments of the subject matter described in this specification can be implemented as one or more computer programs, i.e., one or more modules in computer program instructions encoded on a tangible non-transitory program carrier to be executed by a data processing apparatus or to control the operation of a data processing apparatus. Alternatively or additionally, the program instructions can be encoded on a manually generated propagated signal, such as a machine-generated electrical, optical, or electromagnetic signal, which is generated to encode and transmit information to a suitable receiver apparatus for execution by a data processing apparatus. A computer storage medium can be a machine-readable storage device, a machine-readable storage substrate, a random or serial access memory device, or a combination of one or more of them.

[0104] The processes and logical flows described in this specification can be performed by one or more programmable computers executing one or more computer programs to perform the corresponding functions by operating on input data and generating output. The processes and logical flows can also be performed by, for example, FPGA (Field Programmable Gate Array) or ASIC (Application Specific Integrated Circuit), and the apparatus can also be implemented as dedicated logic circuitry.

[0105] Computers suitable for executing computer programs include, for example, general and / or special-purpose microprocessors, or any other type of central processing unit. Generally, the central processing unit will receive instructions and data from a read-only memory and / or a random access memory. The basic components of a computer include a central processing unit for implementing or executing instructions and one or more memory devices for storing instructions and data. Generally, a computer will also include one or more mass storage devices for storing data, such as magnetic disks, magneto-optical disks, or optical disks, etc., or the computer will be operatively coupled to such mass storage devices to receive data from them or to transmit data to them, or both. However, a computer is not necessarily required to have such devices. In addition, a computer can be embedded in another device, such as a mobile phone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a global positioning system (GPS) receiver, or a portable storage device such as a universal serial bus (USB) flash drive, to name just a few.

[0106] Computer-readable media suitable for storing computer program instructions and data include all forms of non-volatile memory, media, and memory devices, such as semiconductor memory devices (e.g., EPROM, EEPROM, and flash memory devices), magnetic disks (e.g., internal hard disks or removable disks), magneto-optical disks, and CD-ROM and DVD-ROM disks. The processor and the memory may be supplemented by, or incorporated in, special purpose logic circuitry.

[0107] Although this specification contains many specific implementation details, these should not be construed as limiting the scope of any invention or the scope of what is claimed, but rather as mainly describing the features of specific embodiments of a particular invention. Certain features that are described in multiple embodiments in this specification may also be implemented in combination in a single embodiment. On the other hand, the various features described in a single embodiment may also be implemented separately in multiple embodiments or in any suitable sub-combination. Additionally, although features may operate in certain combinations as described above and are even initially claimed as such, one or more features from a claimed combination may in some cases be removed from the combination, and the claimed combination may be directed to a sub-combination or a variation of a sub-combination.

[0108] Similarly, although the operations are depicted in the drawings in a particular order, this should not be understood as requiring that the operations be performed in the particular order shown or sequentially, or that all of the illustrated operations be performed, to achieve the desired result. In some cases, multitasking and parallel processing may be advantageous. Additionally, the separation of the various system modules and components in the above embodiments should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

[0109] Thus, specific embodiments of the subject matter have been described. Other embodiments are within the scope of the appended claims. In some cases, the acts recited in the claims may be performed in a different order and still achieve the desired result. Additionally, the processes depicted in the drawings are not necessarily in the particular order or sequential order shown to achieve the desired result. In some implementations, multitasking and parallel processing may be advantageous.

[0110] The above are the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as within the protection scope of the present invention.

Claims

1. A precursor data fault diagnosis method based on deep learning, characterized in that: include: Step S1, judging whether it is a vibration high alarm fault according to the average value of the vibration signal collected by the sensor in a certain second; if it is a vibration high alarm fault, the algorithm stops, otherwise go to step S2; Step S2, judging whether there is a gear wear fault trend according to the effective value, peak-to-peak value, one times meshing frequency amplitude, two times meshing frequency amplitude and three times meshing frequency amplitude of the vibration signal; If it is a gear wear fault trend, go to step S5; Otherwise go to step S3; Step S3, judging whether there is a rotor unbalance fault trend according to the effective value, peak value, kurtosis, single frequency amplitude in the X direction, single frequency amplitude in the Y direction, single frequency amplitude, double frequency amplitude and triple frequency amplitude of the vibration signal in the X or Y direction; if there is a rotor unbalance fault trend, go to step S5; Otherwise go to step S4; Step S4, judging whether there is a bearing misalignment fault trend according to the effective value, peak-to-peak value and double speed frequency amplitude of the vibration displacement of the vibration signal at the axis +45 degrees or the axis -45 degrees; If it is a bearing misalignment fault trend, go to step S5; Step S5: According to the vibration signals collected in the certain second and the previous second, a deep learning network based on the VGG-11 network is used to obtain the fault type.

2. The method for predictive data fault diagnosis based on deep learning according to claim 1, characterized in that: In step S1, judging whether it is a vibration high alarm fault according to the average value of the vibration signal collected by the sensor in a certain second includes: If the average value of the vibration signal collected by the sensor in a certain second is greater than the upper limit of the vibration signal, it is determined to be a vibration high alarm fault.

3. The precursor data fault diagnosis method based on deep learning according to claim 1 is characterized in that: In step S2, judging whether there is a gear wear fault trend according to the effective value, peak-to-peak value, one times meshing frequency amplitude, two times meshing frequency amplitude and three times meshing frequency amplitude of the vibration signal includes: If the effective value of the vibration signal > the upper limit of the vibration signal and the peak-to-peak value of the vibration signal > the upper limit of the vibration signal and the one times the meshing frequency amplitude, the two times the meshing frequency amplitude and the three times the meshing frequency amplitude of the vibration signal > the upper limit of the meshing frequency amplitude are true, it is judged as a gear wear fault trend.

4. The precursor data fault diagnosis method based on deep learning according to claim 1 is characterized in that: In step S3, judging whether there is a rotor imbalance fault trend according to the effective value, peak value, kurtosis, single frequency amplitude in the X direction, single frequency amplitude in the Y direction, single frequency amplitude, double frequency amplitude and triple frequency amplitude of the vibration signal in the X or Y direction includes: If the effective value of the vibration signal in the X and Y directions / the peak value of the vibration signal in the X and Y directions>the upper limit of the ratioand the kurtosis>the upper limit of the kurtosisand the single frequency amplitude of the vibration signal in the X direction>the single frequency amplitude in the Y directionandthe single frequency amplitude of the vibration signal in the X or Y direction>the double frequency amplitude of the vibration signal in the X or Y directionandthe single frequency amplitude of the vibration signal in the X or Y direction>the triple frequency amplitude of the vibration signal in the X or Y directionandthe single frequency amplitude of the vibration signal in the X or Y direction>the upper limit of the frequency amplitude is true, it is judged as a rotor imbalance fault trend.

5. The method for fault diagnosis based on precursor data of deep learning according to claim 1, characterized in that: In step S4, judging whether it is a bearing misalignment fault trend according to the effective value, peak-to-peak value and double speed frequency amplitude of the vibration displacement of the axis +45 degrees or the axis -45 degrees of the vibration signal includes: If the effective value of the vibration displacement of the vibration signal at +45 degrees or -45 degrees on the axis > the upper limit of the vibration displacement and the peak-to-peak value of the vibration displacement > the upper limit of the vibration displacement and the amplitude of the twice speed frequency of the vibration displacement > the upper limit of the vibration displacement are true, it is judged as a bearing misalignment fault trend.

6. The method for predictive data fault diagnosis based on deep learning according to claim 1, characterized in that: In step S4, the fault types obtained by using a deep learning network based on a VGG-11 network according to the vibration signals collected in the one second and the previous second include: Draw a time domain curve according to the vibration signals collected in the certain second and the second before it; Subtracting the average value of the vibration signals collected at a certain second and the second before it from the vibration signals collected at a certain second and the second before it, to obtain a transformed vibration signal; Performing FFT transformation on the transformed vibration signal to obtain a frequency spectrum curve; The time domain curve and the spectrum curve are input into a deep learning network based on the VGG-11 network for classification to obtain the fault type.

7. The method for predictive data fault diagnosis based on deep learning according to claim 1, characterized in that: In step S4, the deep learning network based on the VGG-11 network is: Replace the max pooling with average pooling in the VGG-11 network.

8. A precursor data fault diagnosis system based on deep learning, characterized in that: The system comprises: The first processing module is configured to determine whether it is a vibration high alarm fault according to the average value of the vibration signal collected by the sensor in a certain second; if it is a vibration high alarm fault, the system stops judging, otherwise it goes to the second processing module; The second processing module is configured to determine whether there is a gear wear fault trend according to the effective value, peak-to-peak value, one times meshing frequency amplitude, two times meshing frequency amplitude and three times meshing frequency amplitude of the vibration signal; if there is a gear wear fault trend, go to the fifth processing module; otherwise go to the third processing module; The third processing module is configured to determine whether there is a rotor imbalance fault trend according to the effective value, peak value, kurtosis, single frequency amplitude in the X direction, single frequency amplitude in the Y direction, single frequency amplitude, double frequency amplitude and triple frequency amplitude of the vibration signal in the X or Y direction; if there is a rotor imbalance fault trend, turn to the fifth processing module; otherwise, turn to the fourth processing module; The fourth processing module is configured to determine whether there is a bearing misalignment fault trend according to the effective value, peak-to-peak value and double speed frequency amplitude of the vibration displacement of the axis +45 degrees or axis -45 degrees of the vibration signal; if there is a bearing misalignment fault trend, then go to the fifth processing module; The fifth processing module is configured to obtain the fault type by using a deep learning network based on the VGG-11 network according to the vibration signals collected in the certain second and the previous second.

9. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the steps in the precursor data fault diagnosis method based on deep learning described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the steps in the precursor data fault diagnosis method based on deep learning described in any one of claims 1 to 7 are implemented.

Images