CN113125135A - Fault diagnosis method for rotary machine, storage medium, and electronic device - Google Patents
Fault diagnosis method for rotary machine, storage medium, and electronic device Download PDFInfo
- Publication number
- CN113125135A CN113125135A CN202110350245.7A CN202110350245A CN113125135A CN 113125135 A CN113125135 A CN 113125135A CN 202110350245 A CN202110350245 A CN 202110350245A CN 113125135 A CN113125135 A CN 113125135A
- Authority
- CN
- China
- Prior art keywords
- fault
- vibration signal
- fault diagnosis
- vibration
- rotary machine
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 55
- 238000003745 diagnosis Methods 0.000 title claims abstract description 53
- 238000003860 storage Methods 0.000 title claims abstract description 17
- 238000004458 analytical method Methods 0.000 claims abstract description 39
- 238000013528 artificial neural network Methods 0.000 claims abstract description 32
- 238000001228 spectrum Methods 0.000 claims abstract description 20
- 230000003068 static effect Effects 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 12
- 238000006243 chemical reaction Methods 0.000 claims description 11
- 238000000605 extraction Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 10
- 238000007781 pre-processing Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000009432 framing Methods 0.000 claims description 4
- 238000013144 data compression Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 7
- 230000003595 spectral effect Effects 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 6
- 230000008859 change Effects 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 5
- 238000013527 convolutional neural network Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000007726 management method Methods 0.000 description 3
- 230000035945 sensitivity Effects 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012805 post-processing Methods 0.000 description 2
- 230000005236 sound signal Effects 0.000 description 2
- 108010076504 Protein Sorting Signals Proteins 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 210000005069 ears Anatomy 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
- G01H17/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves, not provided for in the preceding groups
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Abstract
The invention provides a fault diagnosis method for a rotary machine, a storage medium and an electronic device, wherein the fault diagnosis method for the rotary machine comprises the following steps: acquiring a vibration signal of a rotating machine; extracting the characteristics of the vibration signals to generate characteristic data; the characteristic data comprises frequency spectrum information of the vibration signal; inputting the characteristic data into a pre-trained neural network for fault grade analysis; and judging the fault type according to the result of the fault grade analysis, and generating a fault diagnosis result based on the fault type. The invention can realize automatic fault diagnosis, can effectively diagnose and identify weak vibration faults, and achieves the function of fault early warning before the rotary machine has larger faults to influence normal work.
Description
Technical Field
The present invention relates to a fault diagnosis method, and more particularly, to a fault diagnosis method for a rotary machine, a storage medium, and an electronic device, which belong to the technical field of fault diagnosis.
Background
Compared with shipping ships, the engineering ships have the characteristics of dense equipment configuration, strong specialization, large load change, worse operation condition and the like, and once a fault occurs, the engineering ships have a large influence range and are difficult to accurately position and quickly repair, so that serious economic loss can be brought. In the early stage of ship management, people can judge the existence of some simple faults and provide repairing measures according to the touch of equipment, the feeling of noise state characteristics and the experience of field management personnel. However, with the development of modern production and the progress of science and technology, the structure of modern equipment is more and more complex, the function is more and more perfect, the automation degree is higher and more, and the existence of mechanical faults is artificially judged only by means of human senses and experiences and cannot meet the requirements of the development of the mechanical automation technology of ships. Therefore, it is important to analyze the failure of the ship mechanical equipment by automated means such as vibration detection, machine learning, and artificial intelligence techniques.
However, because the vibration waveform of the equipment fault is relatively complex, the existing detection method needs to know the relevant information in the vibration waveform signal of the equipment in practical application, clarify the relation between the relevant information and the vibration waveform signal so as to enable the fault characteristic to become clear, and then grasp the spectral line with higher amplitude to analyze the possible factors of the frequency components. In addition, a special field expert is needed for analysis, and the method takes half a day or even longer, so that the cost is high, or a plurality of existing methods often need multiple measurements to accurately grasp the fault type.
Therefore, how to provide a fault diagnosis method for a rotary machine, a storage medium, and an electronic device to solve the defects that the prior art cannot accurately identify the weak fault type of the rotary machine, becomes a technical problem to be solved by those skilled in the art.
Disclosure of Invention
In view of the above-mentioned drawbacks of the prior art, an object of the present invention is to provide a fault diagnosis method for a rotary machine, a storage medium, and an electronic apparatus, which are used to solve the problem that the prior art cannot accurately identify a weak fault type of the rotary machine.
To achieve the above and other related objects, an aspect of the present invention provides a fault diagnosis method for a rotary machine, including: acquiring a vibration signal of a rotating machine; extracting the characteristics of the vibration signals to generate characteristic data; the characteristic data comprises frequency spectrum information of the vibration signal; inputting the characteristic data into a pre-trained neural network for fault grade analysis; and judging the fault type according to the result of the fault grade analysis, and generating a fault diagnosis result based on the fault type.
In an embodiment of the invention, the feature data includes static feature data and dynamic feature data.
In an embodiment of the present invention, the vibration signal is subjected to feature extraction to generate feature data; the step of the characteristic data containing the frequency spectrum information of the vibration signal comprises: preprocessing the vibration signal; performing time-frequency conversion on the preprocessed vibration signals, and performing energy analysis according to the time-frequency conversion result; determining the static feature data and the dynamic feature data based on a result of the energy analysis.
In an embodiment of the present invention, the step of preprocessing the vibration signal includes: pre-emphasis is carried out on the vibration signal, so that the vibration signal obtains a frequency spectrum by using the same signal-to-noise ratio from the whole frequency band from low frequency to high frequency; performing framing processing on the vibration signal to segment the vibration signal with the signal length exceeding a preset length; and windowing the vibration signals so as to ensure that the two ends of each frame of vibration signal ensure the continuity of the signals.
In an embodiment of the present invention, the step of performing time-frequency conversion on the preprocessed vibration signal and performing energy analysis according to a time-frequency conversion result includes: carrying out fast Fourier transform on the preprocessed vibration signals to generate frequency domain signals; and calculating the energy of the Mel filter of the frequency domain signal, and performing data compression on the calculation result to determine the Mel frequency cepstrum coefficient.
In an embodiment of the invention, the step of determining the static feature data and the dynamic feature data based on the result of the energy analysis includes: taking the mel-frequency cepstrum coefficients as the static feature data; and taking the first-order difference of the Mel frequency cepstrum coefficients as the dynamic characteristic data.
In an embodiment of the present invention, the step of inputting the feature data into a pre-trained neural network for fault level analysis includes: inputting the characteristic data into a pre-trained neural network to generate a grade recognition result; determining a fault grade according to the grade identification result; the fault classes include weak faults and non-weak faults.
In an embodiment of the present invention, the step of determining a fault type according to a result of the fault level analysis and generating a fault diagnosis result based on the fault type includes: if the fault grade is a weak fault, inputting the characteristic data into an error back propagation neural network to generate a type identification result; determining the fault type according to the type identification result; the fault types include: weak faults of the diesel engine, the gear box, the motor, the bearing box and the pump.
To achieve the above and other related objects, another aspect of the present invention provides a computer-readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the method for diagnosing a fault of a rotary machine.
To achieve the above and other related objects, a further aspect of the present invention provides an electronic device, comprising: a processor and a memory; the memory is used for storing a computer program, and the processor is used for executing the computer program stored by the memory, so that the electronic equipment executes the fault diagnosis method of the rotating machine.
As described above, the method for diagnosing a failure of a rotary machine, the storage medium, and the electronic apparatus according to the present invention have the following advantageous effects:
(1) compared with the traditional engineering ship vibration fault diagnosis method, the method avoids a post-processing mode, namely judging the fault type after the fault occurs and positioning the fault part, and learns the data characteristics of various faults under normal working conditions and micro-abnormal working conditions by combining MFCC (Mel-Frequency cepstrum Coefficients or Mel Frequency cepstrum Coefficients) characteristic extraction and a deep neural network, thereby effectively diagnosing the weak vibration fault and achieving the function of fault early warning.
(2) When weak faults are detected, the type of the weak faults is automatically analyzed by combining a traditional time domain and frequency domain analysis method, so that fault early warning is carried out on the specific weak fault type.
(3) Compared with the traditional fault analysis method based on vibration characteristics, the method can realize automatic fault diagnosis, and the diagnosis result is the fault type and the corresponding probability.
(4) The invention can carry out fault diagnosis only by detecting the vibration signal on the surface of the equipment of the rotating machinery, does not need to change the original structure of the equipment, does not need to disassemble and assemble a machine, and is very convenient to use. In addition, centralized equipment fault monitoring can be carried out through the Internet of things.
Drawings
Fig. 1 is a schematic flow chart illustrating a fault diagnosis method for a rotary machine according to an embodiment of the present invention.
Fig. 2 is a flow chart illustrating feature extraction of a fault diagnosis method for a rotary machine according to an embodiment of the present invention.
Fig. 3 is a schematic diagram illustrating a fault analysis of the method for diagnosing a fault of a rotating machine according to an embodiment of the present invention.
Fig. 4 is a flowchart illustrating a fault diagnosis method of a rotary machine according to an embodiment of the present invention.
Fig. 5 is a schematic structural connection diagram of an electronic device according to an embodiment of the invention.
Description of the element reference numerals
5 electronic device
51 processor
52 memory
S11-S14
S121 to S123
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention. It is to be noted that the features in the following embodiments and examples may be combined with each other without conflict.
It should be noted that the drawings provided in the following embodiments are only for illustrating the basic idea of the present invention, and the drawings only show the components related to the present invention rather than the number, shape and size of the components in actual implementation, and the type, quantity and proportion of the components in actual implementation may be changed freely, and the layout of the components may be more complicated.
The fault diagnosis method, the storage medium and the electronic equipment of the rotary machine can realize automatic fault diagnosis, can effectively diagnose and identify weak vibration faults, and achieve the function of fault early warning before the rotary machine has larger faults to influence normal work. The artificial intelligence technology is applied to the vibration fault diagnosis of the engineering ship rotating machinery, and the detection precision and accuracy are higher. Aiming at the problem that fault data are difficult to obtain, the method for early warning the fault by adopting the deep neural network is provided, only normal operation data and data in slight abnormality of equipment are required to be collected, and compared with the method for collecting various fault data, the development cost is lower. The voice signal processing MFCC technology is applied to the vibration fault diagnosis of the rotary machinery of the engineering ship, compared with a traditional feature extraction method, the voice signal processing MFCC technology can better extract fault features, and therefore the recognition accuracy is greatly improved.
The principles and embodiments of a fault diagnosis method, a storage medium, and an electronic device of a rotary machine according to the present embodiment will be described in detail below with reference to fig. 1 to 5, so that those skilled in the art can understand the fault diagnosis method, the storage medium, and the electronic device of the rotary machine according to the present embodiment without creative work.
Referring to fig. 1, a schematic flow chart of a fault diagnosis method for a rotating machine according to an embodiment of the present invention is shown. As shown in fig. 1, the fault diagnosis method for a rotary machine specifically includes the following steps:
s11, a vibration signal of the rotary machine is acquired.
In one embodiment, a vibration sensor is used to obtain a vibration signal of the rotating machine from a surface of the rotating machine. The vibration signal is a waveform signal such as an audio signal or an electric signal which is generated according to the conversion of mechanical quantity in the vibration process of the rotating machine and is proportional to the mechanical quantity. Specifically, the surface vibration of rotary machines such as oil dispensers, cabin pumps, supercharger bearings, auxiliary blowers, and intermediate bearings may be tested to obtain the vibration signal.
Further, the vibration sensor can transmit vibration signals to equipment for feature extraction and analysis in a wireless data transmission mode, so that the concentrated weak fault monitoring of the rotary machine through the Internet of things is realized. In practical applications, the wireless data transmission mode may be WIFI, Radio, Zigbee (Zigbee), GPRS (General Packet Radio Service), or other transmission modes capable of constructing a data network of the internet of things.
In practical application, different mechanical equipment of a ship has different vibration fault characteristics, and common rotating mechanical equipment and common fault characteristics thereof are subjected to list management to form a fault type table shown in table 1.
TABLE 1 Fault types Table
As can be seen from table 1, different types of rotating machines in the engineering ship have different vibration faults, and before the vibration fault in table 1 occurs, the vibration signal of the rotating machine has already shown an abnormality in advance.
S12, extracting the characteristics of the vibration signals to generate characteristic data; the characteristic data includes frequency spectrum information of the vibration signal.
In one embodiment, the feature data includes static feature data and dynamic feature data.
Specifically, the collected vibration signals are subjected to feature extraction by adopting an MFCC analysis method which is most widely applied in the aspects of voice recognition and analysis. MFCC is based on the auditory characteristics of people, and the linear frequency spectrum is corresponding to the Mel frequency nonlinear frequency spectrum, and finally, the cepstrum transformation is carried out.
Referring to fig. 2, a feature extraction flow chart of a fault diagnosis method for a rotating machine according to an embodiment of the present invention is shown.
As shown in fig. 2, S12 includes the steps of:
and S121, preprocessing the vibration signal.
In one embodiment, S121 includes the following steps:
(1) and pre-emphasizing the vibration signal so as to obtain a frequency spectrum by using the same signal-to-noise ratio in the whole frequency band from low frequency to high frequency of the vibration signal.
Specifically, the high-frequency part of the signal is improved by using the high-pass filter, so that the frequency spectrum is flatter, and the frequency spectrum can be obtained by using the same signal-to-noise ratio in the whole frequency band from low frequency to high frequency, which is the main purpose of pre-emphasis. The transfer function of the high-pass filter is:
y(n)＝x(n)-αx(n-1)
where x (n) is the original signal, y (n) is the emphasized signal sequence, and α is the pre-emphasis coefficient.
(2) And performing framing processing on the vibration signal to segment the vibration signal with the signal length exceeding a preset length.
In particular, framing is the division of a longer vibration signal into small segments. The time covered by one frame is 20-30 ms, wherein an overlap region is formed between two adjacent frames, so as to avoid the excessive change of the two adjacent frames. For example, the invention takes 384 points as one frame, and the length of the selected overlapping area is one half frame, i.e. 192 points.
(3) And windowing the vibration signals so as to ensure that the two ends of each frame of vibration signal ensure the continuity of the signals.
In particular, windowing is performed to avoid a situation where signal discontinuities may occur at both ends of the vibration signal per frame. Commonly used window functions include square windows, Hamming windows, Hanning windows, and other functions that can implement windowing. In contrast, the Hamming window is suitable for processing signals with complex frequency spectrum representation and a plurality of frequency spectrum components, and the Hamming window is preferably added to each frame of signal.
And S122, performing time-frequency conversion on the preprocessed vibration signals, and performing energy analysis according to the time-frequency conversion result.
In one embodiment, S122 includes the following steps:
(1) and carrying out fast Fourier transform on the preprocessed vibration signals to generate frequency domain signals.
In particular, it is difficult to analyze the characteristics of the vibration signal by observing the transformation of the signal in the time domain, so the signal is usually time-frequency transformed by Fast Fourier Transform (FFT) to observe the energy distribution. As shown in the following formula:
X(i,k)＝FFT[xi(n)]
wherein x isiAnd (n) is an original time domain signal of the ith frame, and X (i, k) is a frequency domain signal of the kth spectral line of the ith frame.
(2) And calculating the energy of the Mel filter of the frequency domain signal, and performing data compression on the calculation result to determine the Mel frequency cepstrum coefficient. The method specifically comprises the following steps:
(2-1) calculating a Mel filter (Mel filter) energy of the frequency domain signal.
Since the human ear has different sensitivity degrees to different frequency signals, and the relationship is nonlinear, the frequency spectrum is divided into a plurality of Mel filter banks according to the sensitivity degrees of different human ears, and then the spectral line energy of each frame of FFT data is calculated, as shown in the following formula:
E(i,k)＝[X(i,k)]2
wherein, X (i, k) is the frequency domain signal of the kth spectral line of the ith frame, E (i, k) is the energy spectrum of the sound signal, wherein i represents the ith frame, and k represents the kth spectral line. Then the spectral energy of each frame is passed through Mel filter, i.e. the energy spectrum E (i, k) of each frame is multiplied by the frequency domain of Mel filter and added, i.e.
Wherein Hm(k) Representing the mth filter in the mel filter bank, M representing the number of filter banks, e.g., 26, and S (i, M) representing the energy of the ith frame after passing through the mth mel filter.
And (2-2) performing discrete cosine transform on the Mel filter energy.
Discrete Cosine Transform (DCT), which is mainly used for compression, such as compressing data or compressing images, can convert signals in a spatial domain to a frequency domain with good decorrelation. The spatial domain is also called a spatial domain, that is, a pixel domain, and the processing in the spatial domain is processing at a pixel level, such as image superposition at the pixel level. After fourier transformation, the spectrum of the image, i.e. the energy gradient representing the image, is obtained.
Wherein S (i, m) represents the energy of the ith frame after passing through the mth mel filter, n represents the spectral line after discrete cosine transform, and L is the order of the mel-frequency cepstrum coefficient, for example, the value is 18.
S123, determining the static characteristic data and the dynamic characteristic data based on the result of the energy analysis.
In one embodiment, the mel-frequency cepstrum coefficients are used as the static feature data; and taking the first-order difference of the Mel frequency cepstrum coefficients as the dynamic characteristic data.
Specifically, the MFCC coefficients calculated in step (2-2) reflect the static characteristics of the vibration signal, and the dynamic characteristics of the vibration signal can be further acquired by using the dynamic difference spectrum. Therefore, the analysis of the static characteristic data and the dynamic characteristic data can effectively improve the overall identification performance of the system.
And S13, inputting the characteristic data into a pre-trained neural network for fault level analysis. Thereby, a classification of the severity of faults of the rotating machine is achieved, e.g. the severity comprises weak faults and non-weak faults (critical faults and other faults). The main objective is to identify the level of weak failure of the rotating machine based on the grade.
In one embodiment, the feature data is input into a pre-trained neural network to generate a grade recognition result; determining a fault grade according to the grade identification result; the fault classes include weak faults and non-weak faults.
In one embodiment, the pre-trained neural network comprises: the method comprises the steps of obtaining a large number of vibration signals when the rotary machine is slightly abnormal (weak fault) in advance, using the vibration signals as given training data, setting corresponding labels for the given training data (namely setting which weak fault type the given training data specifically corresponds to), and finishing training of the deep neural network according to the characteristics of the automatically extracted vibration signals so as to realize fault early warning by using the trained deep neural network to identify the weak fault.
The deep neural network has the advantages that the deep neural network has stronger anti-interference capability because a large amount of vibration signal data are learned by the deep neural network, the robustness shown by the noise is better, and misjudgment caused by noise sensitivity is avoided. Further, the MFCC technology is combined with a deep neural network, so that the defect that the application effect depends on the expression capability of the features in the existing method can be overcome. Therefore, even if the ship mechanical equipment is subjected to more nonlinear influence factors, the difference between the signals can be well expressed through the characteristics, and the learning effect is effectively improved.
Please refer to fig. 3, which is a schematic diagram illustrating a fault analysis of the method for diagnosing a fault of a rotating machine according to an embodiment of the present invention. As shown in fig. 3, the extracted MFCC fault signature is trained using a convolutional neural network. The adopted convolutional neural network has 16 resblocks in total and finally passes through a full connection layer and a Softmax layer.
Specifically, the characteristic data of the vibration signal is input into 16 reblock units, and then input into the full connection layer and the Softmax layer for identification, and the identification result and the probability of whether the weak fault occurs are output.
It should be noted that, the adoption of 16 resblocks is only one of the embodiments of the convolutional neural network in the present invention, and a convolutional neural network formed by resblocks of other numbers than 16 resblocks is also within the protection scope of the present invention, and the larger the resblock number is, the larger the calculation amount is, the slower the speed is, but the implementation of the fault analysis principle in the present invention is not affected. Resblock is a unit of ResNet (Residual Network).
And S14, judging the fault type according to the result of the fault level analysis, and generating a fault diagnosis result based on the fault type.
In one embodiment, if the fault level is a weak fault, the feature data is input into an error back propagation neural network to generate a type identification result; determining the fault type according to the type identification result; the fault type corresponds to a serious fault type in variable 1, and comprises the following steps: weak faults of the diesel engine, the gear box, the motor, the bearing box and the pump.
It should be noted that the error back propagation neural network is only one embodiment of the fault type determination, and other analysis methods and neural networks that can perform time domain and/or frequency domain analysis on the vibration signal are also within the scope of the present invention.
In another embodiment, if the fault level is a weak fault, the feature data may be further input to the neural network for further analysis to generate a type identification result; and determining the fault type according to the type identification result. The neural network is a deep learning network.
Referring to fig. 4, a fault diagnosis flowchart of a fault diagnosis method of a rotating machine according to an embodiment of the present invention is shown. As shown in fig. 4, in an embodiment, the fault level is identified by a deep Neural Network, and further, when the fault level is determined to be a weak fault, the type of the weak fault is specifically identified by a Back Propagation Neural Network (BP Neural Network). The method specifically comprises the following steps:
acquiring a vibration signal of a rotary machine, converting the vibration signal into an MFCC (Mel frequency cepstrum coefficient) characteristic diagram, inputting the characteristic diagram into a deep neural network, identifying whether a weak fault occurs or not through the deep neural network, and identifying the type of the weak fault through a BP (back propagation) neural network if the weak fault occurs; and if not, analyzing the subsequently acquired vibration signal. The MFCC characteristic diagram refers to map information formed after the vibration signal is subjected to characteristic extraction by using an MFCC analysis method, namely characteristic data.
The protection scope of the fault diagnosis method for a rotary machine according to the present invention is not limited to the execution sequence of the steps illustrated in the embodiment, and all the solutions of the prior art including the steps addition, subtraction, and step replacement according to the principle of the present invention are included in the protection scope of the present invention.
The present embodiment provides a computer-readable storage medium on which a computer program is stored, which, when executed by a processor, implements the fault diagnosis method of a rotary machine.
Those of ordinary skill in the art will understand that: all or part of the steps for implementing the above method embodiments may be performed by hardware associated with a computer program. The aforementioned computer program may be stored in a computer readable storage medium. When executed, the program performs steps comprising the method embodiments described above; and the aforementioned computer-readable storage media comprise: various computer storage media that can store program codes, such as ROM, RAM, magnetic or optical disks.
Please refer to fig. 5, which is a schematic structural connection diagram of an electronic device according to an embodiment of the present invention. As shown in fig. 5, the present embodiment provides an electronic device 5, which specifically includes: a processor 51 and a memory 52; the memory 52 is configured to store a computer program, and the processor 51 is configured to execute the computer program stored in the memory 52, so that the electronic device 5 executes each step of the fault diagnosis method for the rotary machine.
The Processor 51 may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the Integrated Circuit may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete gate or transistor logic device, discrete hardware component.
The Memory 52 may include a Random Access Memory (RAM), and may further include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory.
In practice, the electronic device may be a computer including some or all of the components of memory, a memory controller, one or more processing units (CPUs), a peripheral interface, RF circuitry, audio circuitry, speakers, a microphone, an input/output (I/O) subsystem, a display screen, other output or control devices, and an external port; the computer includes, but is not limited to, a Personal computer such as a desktop computer, a notebook computer, a tablet computer, a smart phone, a Personal Digital Assistant (PDA for short), and the like, and the electronic device may also be a server, and the server may be arranged on one or more physical servers according to various factors such as functions, loads, and the like, or may also be a cloud server formed by a distributed or centralized server cluster, which is not limited in this embodiment.
In an embodiment, the electronic device is in communication connection with a user terminal, and the user terminal is configured to receive a fault diagnosis result generated by the electronic device and send an early warning prompt message. The early warning prompt information can be in an acousto-optic form (for example, the user terminal is a smart phone, the smart phone starts a bell or flashes through a flashlight), a user terminal vibration form or a user terminal interface pop window form, and any early warning mode that can draw the attention of the user through the user terminal. The user terminal may be a mobile terminal, such as a laptop, a smartphone, or a tablet computer.
Furthermore, the electronic device is further configured to count the number and time of weak faults occurring in the same rotating machine within a preset time period, count weak fault diagnosis conditions of the rotating machines within all monitoring ranges, and present the number and time of weak faults occurring in the preset time period of each rotating machine and a comparison analysis chart of the weak fault conditions of the rotating machines of the same type in different working areas (for example, the motor 1 in the a working site and the motor 1 in the B working site) to a user, where the presenting mode may be a text, a bar/column/broken line/pie-shaped statistical chart, a voice introduction, and other visualization presenting modes based on statistical data. Therefore, a user can intuitively know the weak fault condition of each rotating machine, and further judge whether to need to carry out maintenance or take other related measures such as monitoring according to the frequency and time of the weak faults.
In summary, compared with the conventional vibration fault diagnosis method for the engineering ship, the fault diagnosis method, the storage medium and the electronic device for the rotary machine avoid a post-processing mode, namely judging the fault type after the fault occurs and positioning the fault part, learn the data characteristics of various faults under normal working conditions and micro-abnormal working conditions by combining MFCC feature extraction and a deep neural network, effectively diagnose the weak vibration fault and achieve the function of fault early warning. When weak faults are detected, the types of the weak faults are automatically analyzed by combining a traditional time domain and frequency domain analysis method, so that fault early warning is carried out on the specific weak fault types. Compared with the traditional fault analysis method based on vibration characteristics, the method can realize automatic fault diagnosis, and the diagnosis result is the fault type and the corresponding probability. The invention can carry out fault diagnosis only by detecting the vibration signal on the surface of the equipment of the rotating machinery, does not need to change the original structure of the equipment, does not need to disassemble and assemble a machine, and is very convenient to use. In addition, centralized equipment fault monitoring can be carried out through the Internet of things. The invention effectively overcomes various defects in the prior art and has high industrial utilization value.
The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or change the above-mentioned embodiments without departing from the spirit and scope of the present invention. Accordingly, it is intended that all equivalent modifications or changes which can be made by those skilled in the art without departing from the spirit and technical spirit of the present invention be covered by the claims of the present invention.
Claims (10)
1. A fault diagnosis method for a rotary machine, characterized by comprising:
acquiring a vibration signal of a rotating machine;
extracting the characteristics of the vibration signals to generate characteristic data; the characteristic data comprises frequency spectrum information of the vibration signal;
inputting the characteristic data into a pre-trained neural network for fault grade analysis;
and judging the fault type according to the result of the fault grade analysis, and generating a fault diagnosis result based on the fault type.
2. The fault diagnosis method for a rotary machine according to claim 1, characterized in that: the feature data includes static feature data and dynamic feature data.
3. The method according to claim 2, wherein the vibration signal is subjected to feature extraction to generate feature data; the step of the characteristic data containing the frequency spectrum information of the vibration signal comprises:
preprocessing the vibration signal;
performing time-frequency conversion on the preprocessed vibration signals, and performing energy analysis according to the time-frequency conversion result;
determining the static feature data and the dynamic feature data based on a result of the energy analysis.
4. The fault diagnosis method for a rotary machine according to claim 3, wherein the step of preprocessing the vibration signal includes:
pre-emphasis is carried out on the vibration signal, so that the vibration signal obtains a frequency spectrum by using the same signal-to-noise ratio from the whole frequency band from low frequency to high frequency;
performing framing processing on the vibration signal to segment the vibration signal with the signal length exceeding a preset length;
and windowing the vibration signals so as to ensure that the two ends of each frame of vibration signal ensure the continuity of the signals.
5. The method according to claim 3, wherein the step of performing time-frequency conversion on the preprocessed vibration signal and performing energy analysis according to the time-frequency conversion result includes:
carrying out fast Fourier transform on the preprocessed vibration signals to generate frequency domain signals;
and calculating the energy of the Mel filter of the frequency domain signal, and performing data compression on the calculation result to determine the Mel frequency cepstrum coefficient.
6. The fault diagnosis method for a rotary machine according to claim 5, wherein the step of determining the static characteristic data and the dynamic characteristic data based on the result of the energy analysis includes:
taking the mel-frequency cepstrum coefficients as the static feature data;
and taking the first-order difference of the Mel frequency cepstrum coefficients as the dynamic characteristic data.
7. The method of claim 1, wherein the step of inputting the characteristic data into a pre-trained neural network for fault level analysis comprises:
inputting the characteristic data into a pre-trained neural network to generate a grade recognition result;
determining a fault grade according to the grade identification result; the fault classes include weak faults and non-weak faults.
8. The fault diagnosis method for a rotary machine according to claim 7, wherein the step of determining a fault type from the result of the fault level analysis and generating a fault diagnosis result based on the fault type includes:
if the fault grade is a weak fault, inputting the characteristic data into an error back propagation neural network to generate a type identification result;
determining the fault type according to the type identification result; the fault types include: weak faults of the diesel engine, the gear box, the motor, the bearing box and the pump.
9. A computer-readable storage medium on which a computer program is stored, characterized in that the computer program, when executed by a processor, implements the fault diagnosis method of a rotary machine according to any one of claims 1 to 8.
10. An electronic device, comprising: a processor and a memory;
the memory is configured to store a computer program, and the processor is configured to execute the computer program stored in the memory to cause the electronic device to execute the fault diagnosis method of a rotary machine according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110350245.7A CN113125135A (en) | 2021-03-31 | 2021-03-31 | Fault diagnosis method for rotary machine, storage medium, and electronic device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110350245.7A CN113125135A (en) | 2021-03-31 | 2021-03-31 | Fault diagnosis method for rotary machine, storage medium, and electronic device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113125135A true CN113125135A (en) | 2021-07-16 |
Family
ID=76774436
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110350245.7A Pending CN113125135A (en) | 2021-03-31 | 2021-03-31 | Fault diagnosis method for rotary machine, storage medium, and electronic device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113125135A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113551765A (en) * | 2021-08-17 | 2021-10-26 | 中冶北方（大连）工程技术有限公司 | Sound spectrum analysis and diagnosis method for equipment fault |
| CN114543983A (en) * | 2022-03-29 | 2022-05-27 | 阿里云计算有限公司 | Vibration signal identification method and device |
| CN114964476A (en) * | 2022-05-27 | 2022-08-30 | 中国石油大学(北京) | Fault diagnosis method, device and equipment for oil and gas pipeline system power equipment |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120330495A1 (en) * | 2011-06-23 | 2012-12-27 | United Technologies Corporation | Mfcc and celp to detect turbine engine faults |
| CN108681747A (en) * | 2018-05-11 | 2018-10-19 | 武汉理工大学 | Rotary machinery fault diagnosis based on deep learning and condition monitoring system and method |
| CN110017991A (en) * | 2019-05-13 | 2019-07-16 | 山东大学 | Rolling bearing fault classification method and system based on spectrum kurtosis and neural network |
| CN110455396A (en) * | 2018-05-07 | 2019-11-15 | 光子瑞利科技（北京）有限公司 | Mechanical breakdown method for detecting vibration and system based on Fibre Optical Sensor |
| CN110940539A (en) * | 2019-12-03 | 2020-03-31 | 桂林理工大学 | Machine equipment fault diagnosis method based on artificial experience and voice recognition |
| CN111044285A (en) * | 2019-11-22 | 2020-04-21 | 军事科学院系统工程研究院军用标准研究中心 | Method for diagnosing faults of mechanical equipment under complex conditions |
| CN112052712A (en) * | 2019-11-07 | 2020-12-08 | 国家电力投资集团有限公司 | Power equipment state monitoring and fault identification method and system |
| CN112149511A (en) * | 2020-08-27 | 2020-12-29 | 深圳市点创科技有限公司 | Method, terminal and device for detecting violation of driver based on neural network |
| CN112509599A (en) * | 2020-10-21 | 2021-03-16 | 中国人民解放军陆军炮兵防空兵学院 | Acoustic spectrum fault analysis and diagnosis method based on BP neural network and Mel cepstrum |
-
2021
- 2021-03-31 CN CN202110350245.7A patent/CN113125135A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120330495A1 (en) * | 2011-06-23 | 2012-12-27 | United Technologies Corporation | Mfcc and celp to detect turbine engine faults |
| CN110455396A (en) * | 2018-05-07 | 2019-11-15 | 光子瑞利科技（北京）有限公司 | Mechanical breakdown method for detecting vibration and system based on Fibre Optical Sensor |
| CN108681747A (en) * | 2018-05-11 | 2018-10-19 | 武汉理工大学 | Rotary machinery fault diagnosis based on deep learning and condition monitoring system and method |
| CN110017991A (en) * | 2019-05-13 | 2019-07-16 | 山东大学 | Rolling bearing fault classification method and system based on spectrum kurtosis and neural network |
| CN112052712A (en) * | 2019-11-07 | 2020-12-08 | 国家电力投资集团有限公司 | Power equipment state monitoring and fault identification method and system |
| CN111044285A (en) * | 2019-11-22 | 2020-04-21 | 军事科学院系统工程研究院军用标准研究中心 | Method for diagnosing faults of mechanical equipment under complex conditions |
| CN110940539A (en) * | 2019-12-03 | 2020-03-31 | 桂林理工大学 | Machine equipment fault diagnosis method based on artificial experience and voice recognition |
| CN112149511A (en) * | 2020-08-27 | 2020-12-29 | 深圳市点创科技有限公司 | Method, terminal and device for detecting violation of driver based on neural network |
| CN112509599A (en) * | 2020-10-21 | 2021-03-16 | 中国人民解放军陆军炮兵防空兵学院 | Acoustic spectrum fault analysis and diagnosis method based on BP neural network and Mel cepstrum |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113551765A (en) * | 2021-08-17 | 2021-10-26 | 中冶北方（大连）工程技术有限公司 | Sound spectrum analysis and diagnosis method for equipment fault |
| CN114543983A (en) * | 2022-03-29 | 2022-05-27 | 阿里云计算有限公司 | Vibration signal identification method and device |
| WO2023185801A1 (en) * | 2022-03-29 | 2023-10-05 | 阿里云计算有限公司 | Vibration signal identification method and apparatus |
| CN114964476A (en) * | 2022-05-27 | 2022-08-30 | 中国石油大学(北京) | Fault diagnosis method, device and equipment for oil and gas pipeline system power equipment |
| CN114964476B (en) * | 2022-05-27 | 2023-08-22 | 中国石油大学(北京) | Fault diagnosis method, device and equipment for oil and gas pipeline system moving equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Manhertz et al. | STFT spectrogram based hybrid evaluation method for rotating machine transient vibration analysis | |
| Sadoughi et al. | Physics-based convolutional neural network for fault diagnosis of rolling element bearings | |
| CN113125135A (en) | Fault diagnosis method for rotary machine, storage medium, and electronic device | |
| Jiang et al. | Bearing fault classification based on convolutional neural network in noise environment | |
| Wu et al. | Continuous wavelet transform technique for fault signal diagnosis of internal combustion engines | |
| Wang et al. | Bearing fault diagnosis method based on Hilbert envelope spectrum and deep belief network | |
| CN110702411A (en) | Residual error network rolling bearing fault diagnosis method based on time-frequency analysis | |
| CN112562724A (en) | Speech quality evaluation model, training evaluation method, system, device, and medium | |
| Lu et al. | CEEMD-assisted bearing degradation assessment using tight clustering | |
| CN112364706A (en) | Small sample bearing fault diagnosis method based on class imbalance | |
| CN112052712B (en) | Power equipment state monitoring and fault identification method and system | |
| CN115238740A (en) | Fault diagnosis method and device based on time-frequency distribution and deep learning | |
| CN114048787B (en) | Method and system for intelligently diagnosing bearing fault in real time based on Attention CNN model | |
| CN114639391A (en) | Mechanical failure prompting method and device, electronic equipment and storage medium | |
| WO2023185801A1 (en) | Vibration signal identification method and apparatus | |
| CN112729825A (en) | Method for constructing bearing fault diagnosis model based on convolution cyclic neural network | |
| CN116611184A (en) | Fault detection method, device and medium for gear box | |
| CN116499742A (en) | Rotary machine fault monitoring method based on ensemble learning | |
| CN116226719A (en) | Bearing fault diagnosis method based on multidimensional steady-state vibration characteristics and related components | |
| Martins et al. | Improved variational mode decomposition for combined imbalance-and-misalignment fault recognition and severity quantification | |
| CN115310490A (en) | Rotating equipment fault analysis method based on multi-domain feature and sensitive feature selection | |
| Hao et al. | New fusion features convolutional neural network with high generalization ability on rolling bearing fault diagnosis | |
| Zhong et al. | Parallel learning attention-guided CNN for signal denoising and mechanical fault diagnosis | |
| CN114689298A (en) | Fault detection method and device for walking part of coal mining machine | |
| CN113744756A (en) | Equipment quality inspection and audio data expansion method and related device, equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210716 |
|
| RJ01 | Rejection of invention patent application after publication |