CN112115802A - Crane slewing mechanism gear fault diagnosis method, system and storage medium - Google Patents

Crane slewing mechanism gear fault diagnosis method, system and storage medium Download PDF

Info

Publication number
CN112115802A
CN112115802A CN202010867982.XA CN202010867982A CN112115802A CN 112115802 A CN112115802 A CN 112115802A CN 202010867982 A CN202010867982 A CN 202010867982A CN 112115802 A CN112115802 A CN 112115802A
Authority
CN
China
Prior art keywords
gear
signal
neural network
vibration signal
fault diagnosis
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
Application number
CN202010867982.XA
Other languages
Chinese (zh)
Inventor
王贡献
徐志海
胡志辉
胡勇
袁建明
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan University of Technology WUT
Original Assignee
Wuhan University of Technology WUT
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Wuhan University of Technology WUT filed Critical Wuhan University of Technology WUT
Priority to CN202010867982.XA priority Critical patent/CN112115802A/en
Publication of CN112115802A publication Critical patent/CN112115802A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G06F2218/04—Denoising
    • 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
    • G01M13/02—Gearings; Transmission mechanisms
    • G01M13/021—Gearings
    • 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
    • G01M13/02—Gearings; Transmission mechanisms
    • G01M13/028—Acoustic or vibration analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/12—Classification; Matching

Abstract

The invention relates to a fault diagnosis method, a system and a computer readable storage medium for a gear of a swing mechanism of a crane, wherein the fault diagnosis method for the gear of the swing mechanism of the crane comprises the following steps: acquiring a gear vibration signal, and preprocessing the gear vibration signal to obtain a sample signal; extracting angular domain features of the sample signals, forming the angular domain features into feature vectors, constructing fault feature spaces by the feature vectors, and acquiring principal components in the fault feature spaces; training a BP neural network according to the main components to obtain a BP neural network classifier for gear fault diagnosis; and re-collecting the gear vibration signal, acquiring a main component corresponding to the gear vibration signal, and inputting the main component into a BP neural network classifier for gear fault diagnosis to obtain a gear fault diagnosis result. The method for diagnosing the fault of the gear of the slewing mechanism of the crane can realize early fault diagnosis of the gear of the slewing mechanism of the crane.

Description

Crane slewing mechanism gear fault diagnosis method, system and storage medium
Technical Field
The invention relates to the technical field of crane fault diagnosis, in particular to a method and a system for diagnosing a fault of a gear of a slewing mechanism of a crane and a computer-readable storage medium.
Background
The slewing mechanism is a key part for connecting a base and an upper rotatable part of the crane, the running state of the slewing mechanism is directly related to normal operation and safe use of the whole machine, the mechanism is frequently and emergently started and braked when working, and various early defects such as early cracks, pitting corrosion and the like can be generated after a slewing bearing gear runs for a period of time due to impact vibration generated by the mechanism. If the early fault can not be found in time, the defects are further expanded during operation, serious faults such as tooth breakage and the like occur, so that the machine is violently vibrated, mechanical equipment is damaged or shut down, huge economic loss is caused, and the personal safety of operators is threatened.
The slewing bearing gear teeth of some domestic port cranes do not have any detection device and depend on manual timing inspection, so that huge potential safety hazards exist; along with the increase of the application range and the use of the crane, the occurrence frequency of gear tooth faults is increased, an effective fault detection protection system is designed, and the important point is that the damage of equipment is reduced as much as possible; the traditional gear diagnosis technology based on vibration test usually sets a fixed rotating speed and a fixed load, assumes that abnormal response only comes from equipment deterioration or failure, and realizes gear state monitoring and fault diagnosis by processing, characteristic extraction and mode identification of gear vibration signals under a stable working condition.
When the slewing bearing gear of the crane runs, the running process of the slewing bearing gear is uniform acceleration-uniform speed-uniform deceleration motion, and the generated signal is a non-stationary signal related to the rotating speed; the impact vibration generated by starting and braking causes the signal characteristics to change violently, the early fault characteristics are very weak and are easy to submerge; the slewing mechanism has non-working time intervals in a working cycle of one-time cargo loading and unloading, and the sampling time of the sensor needs to be synchronous with the working time of the mechanism; the commonly used signal processing modes such as fast Fourier transform, short-time Fourier transform and the like are suitable for stable signals, energy leakage is easily caused by wavelet analysis, false harmonic is generated, and the method is not suitable for early fault diagnosis of the crane slewing mechanism gear.
Disclosure of Invention
In view of the above, there is a need to provide a method, a system and a computer readable storage medium for diagnosing gear failure of a crane slewing mechanism, which are used to solve the problem that the prior art is not suitable for early diagnosis of gear failure of a crane slewing mechanism.
The invention provides a fault diagnosis method for a gear of a slewing mechanism of a crane, which comprises the following steps:
acquiring a gear vibration signal, and preprocessing the gear vibration signal to obtain a sample signal;
extracting angular domain features of the sample signals, forming the angular domain features into feature vectors, constructing fault feature spaces by the feature vectors, and acquiring principal components in the fault feature spaces;
training a BP neural network according to the main components to obtain a BP neural network classifier for gear fault diagnosis;
and re-collecting the gear vibration signal, acquiring a main component corresponding to the gear vibration signal, and inputting the main component into a BP neural network classifier for gear fault diagnosis to obtain a gear fault diagnosis result.
The method comprises the steps of carrying out time domain synchronous average denoising on the gear vibration signal, eliminating a signal component irrelevant to a given frequency to obtain a denoised gear vibration signal, converting a variable speed non-stationary signal in the gear vibration signal into a constant speed quasi-stationary signal, decomposing the converted vibration signal into a series of intrinsic mode components, and selecting a signal with a kurtosis value arranged before a set name in the intrinsic mode components to carry out component reconstruction to obtain a sample signal.
Further, the method includes the steps of carrying out time domain synchronous average denoising on the gear vibration signal, eliminating signal components irrelevant to given frequency, and obtaining the denoised gear vibration signal, specifically including intercepting the gear vibration signal x (T) according to the period T of f (T), obtaining N sections in total, adding corresponding points of the sections, and obtaining the denoised gear vibration signal
Figure BDA0002650276350000031
Wherein, tiIs the ith point in the N-segment gear vibration signal, and N (t) is white noise.
Further, the method for converting the variable speed non-stationary signal in the gear vibration signal into the constant speed quasi-stationary signal specifically comprises the steps of determining a resampling time point, calculating the amplitude corresponding to the gear vibration signal according to the resampling time point, obtaining the angular domain of the gear vibration signal, and converting the variable speed non-stationary signal in the gear vibration signal into the constant speed quasi-stationary signal.
Further, the determining the resampling time point specifically includes, by formula
Figure BDA0002650276350000032
Determining a point in time t of the resampling, wherein b0、b1、b2And k is an interpolation coefficient, and theta is a reference shaft angle.
And further training a BP neural network according to the main components to obtain the BP neural network classifier for gear fault diagnosis, wherein the BP neural network classifier for gear fault diagnosis is obtained by taking the main components as the input of the BP neural network, taking the fault types corresponding to the main components as the output of the BP neural network and training the BP neural network through forward transmission and reverse transmission.
Further, training the BP neural network through forward transmission specifically includes obtaining output of an output layer after forward layer-by-layer processing through connection between nodes in the BP neural network.
Further, training the BP neural network through reverse transmission, specifically comprising reversely transmitting the output of the output layer and the error of ideal output to each layer in the BP neural network layer by layer, and loading the error signal to the connection weight according to the gradient principle to convert the connection weight of the whole BP neural network to the direction of reducing the error.
The invention also provides a system for diagnosing the gear fault of the swing mechanism of the crane, which comprises a signal acquisition and preprocessing module, a main component acquisition module, a classifier acquisition module and a gear fault diagnosis module;
the signal acquisition and preprocessing module is used for acquiring a gear vibration signal and preprocessing the gear vibration signal to obtain a sample signal;
the principal component obtaining module is used for extracting angular domain features of the sample signals, forming the angular domain features into feature vectors, and constructing fault feature spaces by the feature vectors to obtain principal components in the fault feature spaces;
the classifier obtaining module is used for training a BP neural network according to the main components to obtain a BP neural network classifier for gear fault diagnosis;
the gear fault diagnosis module is used for re-collecting gear vibration signals, acquiring main components corresponding to the gear vibration signals, and inputting the main components into a BP neural network classifier for gear fault diagnosis to obtain a gear fault diagnosis result.
The invention also provides a computer readable storage medium, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the method for diagnosing the gear fault of the crane slewing mechanism according to any one of the above technical solutions.
Compared with the prior art, the invention has the beneficial effects that: preprocessing the gear vibration signal by acquiring the gear vibration signal to obtain a sample signal; extracting angular domain features of the sample signals, forming the angular domain features into feature vectors, constructing fault feature spaces by the feature vectors, and acquiring principal components in the fault feature spaces; training a BP neural network according to the main components to obtain a BP neural network classifier for gear fault diagnosis; gear vibration signals are collected again, main components corresponding to the gear vibration signals are obtained, and the main components are input into a BP neural network classifier for gear fault diagnosis to obtain gear fault diagnosis results; the early fault diagnosis of the swing mechanism gear of the crane can be realized.
Drawings
Fig. 1 is a schematic flow chart of a method for diagnosing a gear fault of a swing mechanism of a crane according to embodiment 1 of the present invention;
FIG. 2 is a schematic diagram of an arrangement form of a swing mechanism of a crane provided in embodiment 1 of the present invention;
fig. 3 is a schematic diagram of a BP neural network classifier model provided in embodiment 1 of the present invention.
Reference numerals: 1-an electric motor; 2-a brake; 3-an upper swing structure; 4-big gear ring; 5-a rotary pinion; 6-a computer; 7-piezoelectric acceleration sensor.
Detailed Description
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate preferred embodiments of the invention and together with the description, serve to explain the principles of the invention and not to limit the scope of the invention.
Example 1
The embodiment of the invention provides a fault diagnosis method for a gear of a swing mechanism of a crane, which has a flow diagram as shown in figure 1, and comprises the following steps:
s1, collecting gear vibration signals, and preprocessing the gear vibration signals to obtain sample signals;
s2, extracting angular domain features of the sample signals, forming the angular domain features into feature vectors, constructing a fault feature space by the feature vectors, and acquiring principal components in the fault feature space;
s3, training a BP neural network according to the main components to obtain a BP neural network classifier for gear fault diagnosis;
and S4, gear vibration signals are collected again, main components corresponding to the gear vibration signals are obtained, and the main components are input into a BP neural network classifier for gear fault diagnosis to obtain a gear fault diagnosis result.
In a specific embodiment, the layout of the crane slewing mechanism is schematically illustrated, as shown in fig. 2, a piezoelectric acceleration sensor is mounted on the end face of the output pinion shaft of the slewing bearing, the piezoelectric acceleration sensor can reduce the influence of a transmission path on the obtained vibration signal, the mounting mode is an adsorption mode, the slewing mechanism is started, the piezoelectric acceleration sensor starts to receive signals, the slewing mechanism brakes and stops receiving signals, the sensor obtains a gear fault and a normal vibration signal, the gear fault and the normal vibration signal are used as samples of a training classifier after passing through a preprocessing module, the vibration signal is obtained in real time after the classifier is trained, and the trained classifier is used for gear fault recognition; if early failure such as crack, pitting etc. reaches a certain degree, an alarm is given and the operator stops the equipment.
Preferably, the gear vibration signal is preprocessed to obtain a sample signal, and the method specifically comprises the steps of carrying out time domain synchronous average denoising on the gear vibration signal, eliminating a signal component irrelevant to a given frequency to obtain a denoised gear vibration signal, converting a variable speed non-stationary signal in the gear vibration signal into a constant speed quasi-stationary signal, decomposing the converted vibration signal into a series of intrinsic mode components, and selecting a signal with a kurtosis value arranged before a set name in the intrinsic mode components to carry out component reconstruction to obtain the sample signal;
preferably, the method includes the steps of carrying out time domain synchronous average denoising on the gear vibration signal, eliminating signal components irrelevant to given frequency, and obtaining the denoised gear vibration signal, specifically including intercepting the gear vibration signal x (T) according to the period T of f (T), obtaining N sections in total, adding corresponding points of each section, and obtaining the denoised gear vibration signal
Figure BDA0002650276350000061
Wherein, tiIs the ith point in the vibration signal of the N-segment gear, and N (t) is white noise;
in a specific embodiment, the time domain synchronous averaging can eliminate signal components irrelevant to a given frequency, including noise and irrelevant periodic signals, reduce the influence generated by starting and braking impact, work in a noise environment, improve the analysis signal-to-noise ratio, and perform signal enhancement processing on an averaging result; the relevant frequency is selected as the gear meshing frequency, and noise and other irrelevant periodic signals are filtered; the time domain synchronous average calculation process is as follows,
mechanical signal x (t) with periodic repeatability generated in the operation of rotating machinery or reciprocating machinery, if x (t) is composed of periodic signal f (t) and white noise n (t), namely
x(t)＝f(t)+n(t)
Intercepting the signal x (T) with the period T of f (T), obtaining N sections of signals in total, adding corresponding points of each section of signals, and obtaining the white noise irrelevance
Figure BDA0002650276350000071
For x (t)i) Average to obtain output signal
Figure BDA0002650276350000072
Preferably, the method comprises the steps of converting a variable speed non-stationary signal in a gear vibration signal into a constant speed quasi-stationary signal, specifically, determining a resampling time point, calculating an amplitude corresponding to the gear vibration signal according to the resampling time point, obtaining an angular domain of the gear vibration signal, and converting the variable speed non-stationary signal in the gear vibration signal into the constant speed quasi-stationary signal;
preferably, the determining the resampling time point specifically includes, by formula
Figure BDA0002650276350000073
Determining a point in time t of the resampling, wherein b0、b1、b2The undetermined coefficient is determined, k is an interpolation coefficient, and theta is a reference shaft angle;
in one specific embodiment, the order tracking analysis is used for tracking the rotating speed of the rotating shaft and realizing constant angle incremental sampling, and the variable speed non-stationary signal is converted into a constant speed quasi-stationary signal;
in order to determine the resampling time point, an angular acceleration pattern of the reference axis is set, the reference axis is generally considered to make uniform angular acceleration motion in a small time period, and the reference axis angle θ can be expressed as
θ(t)＝b0+b1t+b2t2
In the formula b0、b1、b2For the undetermined coefficients, three pulse time points (t) arriving in sequence are determined1,t2,t3) And the increment of the rotation angle delta phi, i.e.
Figure BDA0002650276350000081
B is obtainediThen, the time of the corresponding corner change can be obtained by solving
Figure BDA0002650276350000082
Wherein k is an interpolation coefficient
Figure BDA0002650276350000083
According to the obtained time point, the amplitude corresponding to the vibration signal can be obtained by utilizing an interpolation algorithm;
in one embodiment, the signal is decomposed into a series of eigenmode components by an Ensemble Empirical Mode Decomposition (EEMD), each eigenmode component is an amplitude modulated signal, wherein the eigenmode components in the high frequency range correspond to each frequency family of the gear vibration signal, and the other eigenmode components are noise; the kurtosis is sensitive to impact signals, the kurtosis value is increased along with the occurrence of faults, and fault information can be enhanced by selecting components with large kurtosis values to reconstruct the signals; selecting a signal with the kurtosis value arranged before a set ranking (such as the 5 th ranking) in the intrinsic mode components for component reconstruction;
selecting signal components with large kurtosis for reconstruction, acquiring characteristics of an angular domain, constructing a characteristic vector space, and performing principal component analysis, wherein the characteristic parameters of the angular domain are common parameters, such as a peak value, a root-mean-square value, a mean value, a variance, a peak factor, the kurtosis, a waveform factor, a pulse index, a margin coefficient, skewness and the like; for constructing a feature vector space, the dimension of the feature parameter is the number of the parameters; the principal component analysis selects the principal components, which is a method for processing high-dimensional characteristic data, the most important characteristics are reserved for the high-dimensional data, unimportant characteristics are removed, and the data processing speed is improved; the processing flow of principal component analysis comprises that the feature vector X is set as { X ═ X1,x2,x3,...,xnMeans are removed to obtain
Figure BDA0002650276350000084
Acquiring the eigenvalue and the eigenvector of C, selecting K eigenvectors with larger eigenvalues as row vectors to form a matrix P, and obtaining a principal component Y which is PX;
preferably, training a BP neural network according to the main components to obtain a BP neural network classifier for gear fault diagnosis, specifically including training the BP neural network by using the main components as input of the BP neural network and using fault types corresponding to the main components as output of the BP neural network through forward transmission and reverse transmission to obtain the BP neural network classifier for gear fault diagnosis;
preferably, the training of the BP neural network is performed through forward transmission, specifically including performing forward layer-by-layer processing through connection between nodes in the BP neural network to obtain output of an output layer;
preferably, the BP neural network is trained through reverse transmission, specifically including reversely transmitting the output of the output layer and the error of ideal output layer by layer to each layer in the BP neural network, and loading an error signal to a connection weight according to a gradient principle, so that the connection weight of the whole BP neural network is converted to a direction of reducing the error;
in a specific embodiment, the principal component obtained by principal component analysis is used as the input of a neural network, and the fault type is used as the output to carry out nonlinear mapping, so that the method has strong self-learning and self-adaptive capabilities; taking two-input two-output neural network as an example, a schematic diagram of a BP neural network classifier model, as shown in fig. 3,
the training process of the BP neural network is divided into a forward transmission part and a reverse transmission part; the BP neural network forward transmission is that actual output is obtained after forward layer-by-layer processing is carried out through the connection condition among all nodes;
the hidden layer outputs are
outh1＝f(w1·i1+w2·i2-b1)
outh2＝f(w3·i1+w4·i2-b2)
The output layer outputs
outo1＝f(w5·outh1+w6·outh2-b3)
outo1＝f(w7·outh1+w8·outh2-b4)
Wherein f (x) is 1/(1+ e ^ s-x)，outh1，outh2Outputting for a hidden layer; w is a1,w2,……w8Is a weight factor; i.e. i1、i2As system input, b1、b2、b3、b4Is a threshold factor; outo1、outo2Outputting for the output layer;
BP neural network reverse transmission, namely, the error is reversely transmitted to the previous layers layer by layer, and an error signal is loaded on the connection weight according to the gradient principle, so that the connection weight of the whole neural network is converted to the direction of reducing the error;
according to
Figure BDA0002650276350000104
Calculating error value, and changing weight value and threshold value according to the following formula
Figure BDA0002650276350000101
Figure BDA0002650276350000102
Wherein E istotalIn order to output the total error,
Figure BDA0002650276350000103
is the output layer error, and η is the learning rate.
Example 2
The embodiment of the invention provides a gear fault diagnosis system of a swing mechanism of a crane, which comprises a signal acquisition and preprocessing module, a main component acquisition module, a classifier acquisition module and a gear fault diagnosis module;
the signal acquisition and preprocessing module is used for acquiring a gear vibration signal and preprocessing the gear vibration signal to obtain a sample signal;
the principal component obtaining module is used for extracting angular domain features of the sample signals, forming the angular domain features into feature vectors, and constructing fault feature spaces by the feature vectors to obtain principal components in the fault feature spaces;
the classifier obtaining module is used for training a BP neural network according to the main components to obtain a BP neural network classifier for gear fault diagnosis;
the gear fault diagnosis module is used for re-collecting gear vibration signals, acquiring main components corresponding to the gear vibration signals, and inputting the main components into a BP neural network classifier for gear fault diagnosis to obtain a gear fault diagnosis result.
Example 3
An embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the method for diagnosing gear faults of a crane slewing mechanism as described in embodiment 1.
The invention discloses a method, a system and a computer readable storage medium for diagnosing gear faults of a swing mechanism of a crane, wherein gear vibration signals are acquired and preprocessed to obtain sample signals; extracting angular domain features of the sample signals, forming the angular domain features into feature vectors, constructing fault feature spaces by the feature vectors, and acquiring principal components in the fault feature spaces; training a BP neural network according to the main components to obtain a BP neural network classifier for gear fault diagnosis; gear vibration signals are collected again, main components corresponding to the gear vibration signals are obtained, and the main components are input into a BP neural network classifier for gear fault diagnosis to obtain gear fault diagnosis results; early fault diagnosis of the gear of the swing mechanism of the crane can be realized;
compared with the traditional manual detection method, the technical scheme of the invention directly analyzes the vibration signal without influencing the normal operation of the crane; the sensor is arranged on the slewing bearing pinion in an adsorption mode, so that the influence of a signal transmission path on the obtained vibration signal is greatly reduced; the signal acquisition time of the sensor is synchronous with the working time of the swing mechanism, so that the influence of other components on the vibration signal is reduced; a time domain averaging method obtains a periodic component related to the meshing frequency, and reduces the influence generated by starting and braking impact; the order tracking method extracts information related to the rotating speed in the vibration signal, can eliminate the influence of rotating speed change on the vibration signal, has the function of secondary filtering, and has good adaptability to fault diagnosis of the slewing bearing gear in variable-speed motion; the EEMD method decomposes a non-stationary original vibration signal containing multiple sources into a single-component sub-signal with fault information, is a self-adaptive decomposition method, focuses on the modulation phenomenon near the meshing frequency, has a high signal-to-noise ratio, and is suitable for gear fault detection; compared with the traditional gear fault detection, the fault diagnosis is carried out by utilizing certain time domain indexes such as peak value, root mean square value, kurtosis and the like, the sensitivity of each index to faults is different, one or two characteristic values are used as classified indexes, information loss is easy to occur, the gear state cannot be well determined, more signal information can be reflected by the principal component obtained by utilizing a principal component analysis method, the angular domain comprehensive index is used as the input of a BP neural network classifier, the classifier has higher accuracy, and the acquired signals can be identified in real time after the classifier is trained;
in addition, the technical scheme of the invention can monitor the quality of the wheel teeth in real time, and can immediately send out warning and process in time when early faults such as abrasion, pitting corrosion, cracks and the like occur, so that the possibility of accidents is greatly reduced, the equipment maintenance cost is reduced, and the economic loss and the personnel damage are reduced; compared with the traditional technical scheme, the method has better noise robustness, higher fault recognition rate under various working conditions, higher recognition speed, real-time detection and important significance for ensuring the safety and continuity of the crane slewing mechanism in working.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.

Claims (10)

1. The method for diagnosing the fault of the gear of the swing mechanism of the crane is characterized by comprising the following steps of:
acquiring a gear vibration signal, and preprocessing the gear vibration signal to obtain a sample signal;
extracting angular domain features of the sample signals, forming the angular domain features into feature vectors, constructing fault feature spaces by the feature vectors, and acquiring principal components in the fault feature spaces;
training a BP neural network according to the main components to obtain a BP neural network classifier for gear fault diagnosis;
and re-collecting the gear vibration signal, acquiring a main component corresponding to the gear vibration signal, and inputting the main component into a BP neural network classifier for gear fault diagnosis to obtain a gear fault diagnosis result.
2. The method as claimed in claim 1, wherein the gear vibration signal is preprocessed to obtain a sample signal, and the method comprises the steps of performing time-domain synchronous average denoising on the gear vibration signal, eliminating a signal component irrelevant to a given frequency to obtain a denoised gear vibration signal, converting a variable-speed non-stationary signal in the gear vibration signal into a constant-speed quasi-stationary signal, decomposing the converted vibration signal into a series of eigenmode components, and selecting a signal with a kurtosis value before the given name in the eigenmode components to perform component reconstruction to obtain the sample signal.
3. The method as claimed in claim 2, wherein the time-domain synchronous mean denoising is performed on the gear vibration signal to eliminate the signal component irrelevant to the given frequency, so as to obtain the denoised gear vibration signal, and the method comprises the steps of intercepting the gear vibration signal x (T) with the period T of f (T), obtaining N sections in total, and adding the corresponding points of each section to obtain the denoised gear vibration signal
Figure FDA0002650276340000011
Wherein, tiIs the ith point in the N-segment gear vibration signal, and N (t) is white noise.
4. The method as claimed in claim 2, wherein the step of converting the non-steady-state signal of the gear variation in the gear vibration signal into the quasi-steady-state signal of the constant speed comprises determining a resampling time point, determining an amplitude corresponding to the gear vibration signal according to the resampling time point, obtaining an angular domain of the gear vibration signal, and converting the non-steady-state signal of the gear variation in the gear vibration signal into the quasi-steady-state signal of the constant speed.
5. The method as claimed in claim 2, wherein the determining the resampling time point comprises determining the resampling time point by a formula
Figure FDA0002650276340000021
Determining a point in time t of the resampling, wherein b0、b1、b2And k is an interpolation coefficient, and theta is a reference shaft angle.
6. The method for diagnosing gear faults of a crane slewing mechanism according to claim 1, wherein a BP neural network classifier for diagnosing gear faults is obtained by training a BP neural network according to the main components, and specifically comprises the steps of taking the main components as input of the BP neural network, taking fault types corresponding to the main components as output of the BP neural network, and training the BP neural network through forward transmission and reverse transmission to obtain the BP neural network classifier for diagnosing gear faults.
7. The method for diagnosing gear fault of crane slewing mechanism according to claim 6, wherein the training of the BP neural network is performed by forward transmission, specifically comprising performing forward layer-by-layer processing to obtain the output of the output layer through the connection between each node in the BP neural network.
8. The method as claimed in claim 6, wherein the training of the BP neural network is performed by backward transmission, and specifically comprises the steps of backward transmitting the error between the output of the output layer and the ideal output layer to each layer of the BP neural network layer by layer, and loading the error signal to the connection weight according to the gradient principle, so as to convert the connection weight of the entire BP neural network to the direction of decreasing the error.
9. A fault diagnosis system for a gear of a swing mechanism of a crane is characterized by comprising a signal acquisition and preprocessing module, a main component acquisition module, a classifier acquisition module and a gear fault diagnosis module;
the signal acquisition and preprocessing module is used for acquiring a gear vibration signal and preprocessing the gear vibration signal to obtain a sample signal;
the principal component obtaining module is used for extracting angular domain features of the sample signals, forming the angular domain features into feature vectors, and constructing fault feature spaces by the feature vectors to obtain principal components in the fault feature spaces;
the classifier obtaining module is used for training a BP neural network according to the main components to obtain a BP neural network classifier for gear fault diagnosis;
the gear fault diagnosis module is used for re-collecting gear vibration signals, acquiring main components corresponding to the gear vibration signals, and inputting the main components into a BP neural network classifier for gear fault diagnosis to obtain a gear fault diagnosis result.
10. A computer-readable storage medium, having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method for diagnosing gear failure in a crane slewing mechanism as defined in any one of claims 1 to 8.
CN202010867982.XA 2020-08-26 2020-08-26 Crane slewing mechanism gear fault diagnosis method, system and storage medium Pending CN112115802A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010867982.XA CN112115802A (en) 2020-08-26 2020-08-26 Crane slewing mechanism gear fault diagnosis method, system and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010867982.XA CN112115802A (en) 2020-08-26 2020-08-26 Crane slewing mechanism gear fault diagnosis method, system and storage medium

Publications (1)

Publication Number Publication Date
CN112115802A true CN112115802A (en) 2020-12-22

Family

ID=73804365

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010867982.XA Pending CN112115802A (en) 2020-08-26 2020-08-26 Crane slewing mechanism gear fault diagnosis method, system and storage medium

Country Status (1)

Country Link
CN (1) CN112115802A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112710358A (en) * 2021-03-29 2021-04-27 南京诚远高新科技有限公司 Intelligent machine tool state monitoring device and state monitoring method thereof
CN112839349A (en) * 2020-12-25 2021-05-25 朗坤智慧科技股份有限公司 Angular domain stationary signal denoising method and device based on 5G technology and network server
CN113739909A (en) * 2021-09-08 2021-12-03 上海海事大学 Ship rotating machinery fault diagnosis method based on full-scale time domain averaging

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102434387A (en) * 2011-11-16 2012-05-02 三一电气有限责任公司 Draught fan detection and diagnosis system
CN105606360A (en) * 2015-11-24 2016-05-25 国网内蒙古东部电力有限公司电力科学研究院 Fault diagnosis method for condition-variable planetary gear box based on multi-sensor information fusion
CN106529581A (en) * 2016-10-24 2017-03-22 杭州电子科技大学 Bayesian-network-based bridge type crane fault diagnosis method
CN111521398A (en) * 2020-03-18 2020-08-11 武汉工程大学 Gear box fault diagnosis method and system based on BP neural network and principal component analysis method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102434387A (en) * 2011-11-16 2012-05-02 三一电气有限责任公司 Draught fan detection and diagnosis system
CN105606360A (en) * 2015-11-24 2016-05-25 国网内蒙古东部电力有限公司电力科学研究院 Fault diagnosis method for condition-variable planetary gear box based on multi-sensor information fusion
CN106529581A (en) * 2016-10-24 2017-03-22 杭州电子科技大学 Bayesian-network-based bridge type crane fault diagnosis method
CN111521398A (en) * 2020-03-18 2020-08-11 武汉工程大学 Gear box fault diagnosis method and system based on BP neural network and principal component analysis method

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
仝奇 等: "基于 KPCA 的 BP 神经网络齿轮泵故障诊断方法研究", 《专题技术与工程应用》 *
姜长城: "门式起重机远程监测及诊断分析系统研发_", 《中国优秀硕士论文全文数据库工程科技Ⅱ辑》 *
贺艳 等: "基于 BP 神经网络对减速箱齿轮副的故障诊断与仿真研究", 《机械研究与应用》 *
黄勇 等: "基于深度学习的齿轮泵故障诊断方法研究", 《组合机床与自动化加工技术》 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112839349A (en) * 2020-12-25 2021-05-25 朗坤智慧科技股份有限公司 Angular domain stationary signal denoising method and device based on 5G technology and network server
CN112710358A (en) * 2021-03-29 2021-04-27 南京诚远高新科技有限公司 Intelligent machine tool state monitoring device and state monitoring method thereof
CN112710358B (en) * 2021-03-29 2021-06-25 南京诚远高新科技有限公司 Intelligent machine tool state monitoring device and state monitoring method thereof
CN113739909A (en) * 2021-09-08 2021-12-03 上海海事大学 Ship rotating machinery fault diagnosis method based on full-scale time domain averaging

Similar Documents

Publication Publication Date Title
CN112115802A (en) Crane slewing mechanism gear fault diagnosis method, system and storage medium
Saidi et al. Bi-spectrum based-EMD applied to the non-stationary vibration signals for bearing faults diagnosis
Jayaswal et al. Machine fault signature analysis
Yang et al. Vibration feature extraction techniques for fault diagnosis of rotating machinery: a literature survey
Feng et al. A novel order spectrum-based Vold-Kalman filter bandwidth selection scheme for fault diagnosis of gearbox in offshore wind turbines
Xu et al. Multireceptive field denoising residual convolutional networks for fault diagnosis
CN111259990B (en) Seawater pump fault monitoring device and diagnosis method based on multi-source information fusion
CN109596349A (en) A kind of decelerator trouble diagnostic method based on VMD and PCT
CN102155988A (en) Equipment monitoring and diagnosing method
Alkhadafe et al. Condition monitoring of helical gears using automated selection of features and sensors
Soualhi et al. PHM survey: implementation of signal processing methods for monitoring bearings and gearboxes
Shuiqing et al. Gear fault diagnosis in variable speed condition based on multiscale chirplet path pursuit and linear canonical transform
CN114964769A (en) Wind power gear box vibration signal fault diagnosis method
Zhang et al. Fault diagnosis techniques for planetary gearboxes under variable conditions: A review
Li et al. Instantaneous angular speed-based fault diagnosis of multicylinder marine diesel engine using intrinsic multiscale dispersion entropy
CN107490477A (en) The Fault Diagnosis of Gear Case method compared based on frequency spectrum kernel density function correlation
CN114354187B (en) Gear fault classification detection method and system based on identification of meshing stiffness
Zhou et al. Research on an early warning method for bearing health diagnosis based on EEMD-PCA-ANFIS
Yao et al. A low-frequency fault detection method for low-speed planetary gearbox based on acoustic signals
Yang et al. Intelligent diagnosis technology of wind turbine drive system based on neural network
CN109238728B (en) Method and system for diagnosing faults of parts on vehicle engine
CN115186692A (en) Intelligent fault diagnosis method based on empirical mode decomposition and image conversion
Ma et al. Motor Bearing Fault Diagnosis in an Industrial Robot Under Complex Variable Speed Conditions
Li et al. Self-Iterated Extracting Wavelet Transform and Its Application to Fault Diagnosis of Rotating Machinery
Cheng Self-tuning method of mechanical vibration based on rotation of marine power equipment

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20201222