CN108871760B - Efficient gear fault mode identification method - Google Patents
Efficient gear fault mode identification method Download PDFInfo
- Publication number
- CN108871760B CN108871760B CN201810576763.9A CN201810576763A CN108871760B CN 108871760 B CN108871760 B CN 108871760B CN 201810576763 A CN201810576763 A CN 201810576763A CN 108871760 B CN108871760 B CN 108871760B
- Authority
- CN
- China
- Prior art keywords
- gear
- fault
- calculating
- output
- signal
- 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.)
- Active
Links
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
- 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
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a high-efficiency gear fault mode identification method, which comprises the following steps: firstly, extracting an initial signal, namely acquiring a vibration signal E of a gearbox by adopting a vibration acceleration sensor; secondly, calculating a fault area outline E1, calculating according to the initial signal E, and eliminating weak signals to obtain a fault area outline E1; thirdly, calculating gear contour information E2 and E3, and calculating a gradient value in the belt direction of the contour E1 of the fault area to obtain gear contour information E2 and E3; fourthly, double-threshold calculation is carried out, and double thresholds are automatically calculated according to the fault region profile E1 and the gear profile information E3; and fifthly, restoring the gear image, detecting, filtering and connecting edges of the gear contour information E3 by using a double-threshold method, and restoring to obtain the gear image in the fault state. The efficient gear fault mode identification method has the characteristics of high positioning accuracy, efficient and rapid operation and strong service life prediction reliability.
Description
Technical Field
The invention relates to the technical field of gear fault identification, in particular to a high-efficiency gear fault mode identification method.
Background
The gear is a mechanical transmission form commonly adopted by mechanical equipment, and is very easily damaged and breaks down due to the action of pulsating cyclic stress in the working process, so that a mechanical transmission system fails, and the safe and reliable operation of the whole mechanical equipment is further influenced.
Disclosure of Invention
The invention provides a method for diagnosing the fault of the gear during the whole life from the generation and development of the fault of the gear to the failure, which realizes the identification of the fault modes of the gear at different failure stages and provides decision basis for the optional maintenance and life prediction of the gear box.
The invention can be realized by the following technical scheme:
the invention discloses a high-efficiency gear fault mode identification method, which comprises the following steps:
firstly, extracting an initial signal, namely acquiring a vibration signal E of a gearbox by adopting a vibration acceleration sensor;
secondly, calculating a fault area outline E1, calculating according to the initial signal E, and eliminating weak signals to obtain a fault area outline E1;
thirdly, calculating gear contour information E2 and E3, and calculating a gradient value in the belt direction of the contour E1 of the fault area to obtain gear contour information E2 and E3;
fourthly, double-threshold calculation is carried out, and double thresholds are automatically calculated according to the fault region profile E1 and the gear profile information E3;
and fifthly, restoring the gear image, detecting, filtering and connecting edges of the gear contour information E3 by using a double-threshold method, and restoring to obtain the gear image in the fault state.
Further, the second step of the calculation of the fault area profile E1 includes the following processes: and sorting the data in the initial signal E from small to large according to the numerical values, taking 30% of the numerical values at all the numerical values of 70% as a lowest threshold value, and removing edges smaller than the lowest threshold value from the initial signal E to obtain a fault area profile E1.
Further, the third step of the calculation of the gear profile information E2 and E3 comprises the following steps:
thirdly, the calculation of the gear profile information E2 and E3 comprises the following steps:
the method adopts a square error cost function to discuss various types of problems, wherein the types of the problems are c, the training samples are N, and the error item calculation process is as follows:
in the formula, the first step is that,
denotes a kth dimension of a label corresponding to the nth sample, t denotes a vector of the kth dimension,
a kth output representing a network output corresponding to the nth sample; for multi-class problems, the output is generally organized in a form of "one-of-c", that is, only the output node output of the class corresponding to the input is positive, and the bits or nodes of other classes are 0 or negative, depending on the activation function of the output layer; sigmoid is 0, tanh is-1;
based on the fact that the error over the entire training set is only the sum of the errors of each training sample, the BP for one sample, the error for the nth sample, is expressed as:
then calculating a partial derivative of the cost function E about each weight of the network according to a BP rule; the current layer is represented by l, and the output of the current layer can be represented as:
Xl＝f(ul),with,ul＝Wlxl-1+bl
in the formula, X represents the output, u represents the input, W represents the weight, b represents the offset, where l represents the layer number of the network, and the gear profile information E2, E3 is finally obtained.
Further, the fourth step of the dual threshold calculation comprises the steps of:
d1, sorting the non-zero points in the gear profile information E3 from small to large;
d2, taking the value of L bits in the sequence as the upper threshold, L equals to the number of non-zeros in E3 minus 30% of the number of non-zeros in E1, L ═ len (E3) -len (E1) × 0.3;
d3, take the value of L x0.7 bits in the sequence as the lower threshold.
Further, the fifth step of the gear image restoration comprises the following steps:
e1, filtering the strong signals larger than the upper threshold value, and filtering false strong signals with the signal intensity smaller than an integer k;
e2, connecting the filtered strong signal with the weak signal which is larger than the lower threshold value to obtain the gear image of the fault state.
The efficient gear fault mode identification method has the following beneficial technical effects:
according to the invention, the fault state of the gear is restored by adopting the mathematical model of image recognition enhancement processing, so that the gear fault mode recognition in different failure stages is realized, decision basis is provided for the visual maintenance and the life prediction of the gear box, the image edge of the fault gear can be accurately outlined, false edge information caused by noise and signal conduction can be effectively eliminated, the height threshold value of the fault gear contour edge detection can be adaptively changed, the fault gear contour edge detection effect and the automation degree are improved on the premise of increasing a small amount of operation, and finally the gear fault position is recognized in colleges and universities.
Detailed Description
In order to make the technical solution of the present invention better understood by those skilled in the art, the following detailed description is provided for the product of the present invention with reference to the examples.
The invention discloses a high-efficiency gear fault mode identification method, which comprises the following steps:
firstly, extracting an initial signal, namely acquiring a vibration signal E of a gearbox by adopting a vibration acceleration sensor;
secondly, calculating a fault area outline E1, calculating according to the initial signal E, and eliminating weak signals to obtain a fault area outline E1;
thirdly, calculating gear contour information E2 and E3, and calculating a gradient value in the belt direction of the contour E1 of the fault area to obtain gear contour information E2 and E3;
fourthly, double-threshold calculation is carried out, and double thresholds are automatically calculated according to the fault region profile E1 and the gear profile information E3;
and fifthly, restoring the gear image, detecting, filtering and connecting edges of the gear contour information E3 by using a double-threshold method, and restoring to obtain the gear image in the fault state.
Further, the second step of the calculation of the fault area profile E1 includes the following processes: and sorting the data in the initial signal E from small to large according to the numerical values, taking 30% of the numerical values at all the numerical values of 70% as a lowest threshold value, and removing edges smaller than the lowest threshold value from the initial signal E to obtain a fault area profile E1.
Further, the third step of the calculation of the gear profile information E2 and E3 comprises the following steps:
the method adopts a square error cost function to discuss various types of problems, wherein the types of the problems are c, the training samples are N, and the error item calculation process is as follows:
in the formula, the first step is that,
denotes a kth dimension of a label corresponding to the nth sample, t denotes a vector of the kth dimension,
a kth output representing a network output corresponding to the nth sample; for multi-class problems, the output is generally organized in a form of "one-of-c", that is, only the output node output of the class corresponding to the input is positive, and the bits or nodes of other classes are 0 or negative, depending on the activation function of the output layer; sigmoid is 0, tanh is-1;
based on the fact that the error over the entire training set is only the sum of the errors of each training sample, the BP for one sample, the error for the nth sample, is expressed as:
then calculating a partial derivative of the cost function E about each weight of the network according to a BP rule; the current layer is represented by l, and the output of the current layer can be represented as:
Xl＝f(ul),with,ul＝Wlxl-1+bl
in the formula, X represents the output, u represents the input, W represents the weight, b represents the offset, where l represents the layer number of the network, and the gear profile information E2, E3 is finally obtained.
Further, the fourth step of the dual threshold calculation comprises the steps of:
d1, sorting the non-zero points in the gear profile information E3 from small to large;
d2, taking the value of L bits in the sequence as the upper threshold, L equals to the number of non-zeros in E3 minus 30% of the number of non-zeros in E1, L ═ len (E3) -len (E1) × 0.3;
d3, take the value of L x0.7 bits in the sequence as the lower threshold.
Further, the fifth step of the gear image restoration comprises the following steps:
e1, filtering the strong signals larger than the upper threshold value, and filtering false strong signals with the signal intensity smaller than an integer k;
e2, connecting the filtered strong signal with the weak signal which is larger than the lower threshold value to obtain the gear image of the fault state.
The foregoing is merely a preferred embodiment of the invention and is not intended to limit the invention in any manner; as will be readily apparent to those skilled in the art from the disclosure herein, the present invention may be practiced without these specific details; however, those skilled in the art should appreciate that they can readily use the disclosed conception and specific embodiments as a basis for designing or modifying other structures for carrying out the same purposes of the present invention; meanwhile, any changes, modifications, and evolutions of the equivalent changes of the above embodiments according to the actual techniques of the present invention are still within the protection scope of the technical solution of the present invention.
Claims (3)
1. A high-efficiency gear fault mode identification method is characterized by comprising the following steps:
firstly, extracting an initial signal, namely acquiring a vibration signal E of a gearbox by adopting a vibration acceleration sensor;
secondly, calculating a fault area outline E1, calculating according to the initial signal E, and eliminating weak signals to obtain a fault area outline E1;
thirdly, calculating gear contour information E2 and E3, and calculating a gradient value in the belt direction of the contour E1 of the fault area to obtain gear contour information E2 and E3;
fourthly, double-threshold calculation is carried out, and double thresholds are automatically calculated according to the fault region profile E1 and the gear profile information E3;
fifthly, restoring a gear image, detecting, filtering and connecting edges of gear contour information E3 by using a double-threshold method, and restoring to obtain a gear image in a fault state;
the second step of the calculation of the fault area profile E1 includes the following processes: sorting the data in the initial signal E from small to large according to the numerical values, taking 30% of the numerical values at 70% of all the numerical values as a lowest threshold value, and removing the edge smaller than the lowest threshold value from the initial signal E to obtain a fault area outline E1;
thirdly, the calculation of the gear profile information E2 and E3 comprises the following steps: the method adopts a square error cost function to discuss various types of problems, wherein the types of the problems are c, the training samples are N, and the error item calculation process is as follows:
in the formula, the first step is that,
denotes a kth dimension of a label corresponding to the nth sample, t denotes a vector of the kth dimension,
a kth output representing a network output corresponding to the nth sample; for multi-class problems, the output is generally organized in a form of "one-of-c", that is, only the output node output of the class corresponding to the input is positive, and the bits or nodes of other classes are 0 or negative, depending on the activation function of the output layer; sigmoid is 0, tanh is-1;
based on the fact that the error over the entire training set is only the sum of the errors of each training sample, the BP for one sample, the error for the nth sample, is expressed as:
then calculating a partial derivative of the cost function E about each weight of the network according to a BP rule; the current layer is represented by l, and the output of the current layer can be represented as:
Xl＝f(ul),with,ul＝Wlxl-1+bl
in the formula, X represents the output, u represents the input, W represents the weight, b represents the offset, where l represents the layer number of the network, and the gear profile information E2, E3 is finally obtained.
2. A high efficiency gear failure mode identification method as claimed in claim 1, wherein: the fourth step the dual threshold calculation comprises the steps of:
d1, sorting the non-zero points in the gear profile information E3 from small to large;
d2, taking the value of L bits in the sequence as the upper threshold, L equals to the number of non-zeros in E3 minus 30% of the number of non-zeros in E1, L ═ len (E3) -len (E1) × 0.3;
d3, take the value of L x0.7 bits in the sequence as the lower threshold.
3. A high efficiency gear failure mode identification method as claimed in claim 2, wherein: the fifth step of the gear image restoration comprises the following steps:
e1, filtering the strong signals larger than the upper threshold value, and filtering false strong signals with the signal intensity smaller than an integer k;
e2, connecting the filtered strong signal with the weak signal which is larger than the lower threshold value to obtain the gear image of the fault state.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810576763.9A CN108871760B (en) | 2018-06-07 | 2018-06-07 | Efficient gear fault mode identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810576763.9A CN108871760B (en) | 2018-06-07 | 2018-06-07 | Efficient gear fault mode identification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108871760A CN108871760A (en) | 2018-11-23 |
| CN108871760B true CN108871760B (en) | 2020-07-17 |
Family
ID=64336970
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810576763.9A Active CN108871760B (en) | 2018-06-07 | 2018-06-07 | Efficient gear fault mode identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108871760B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109708877B (en) * | 2018-12-27 | 2020-11-24 | 昆明理工大学 | Mechanical fault analysis method based on wavelet fuzzy recognition and image analysis theory |
| CN113063589B (en) * | 2021-03-18 | 2023-07-04 | 重庆青山工业有限责任公司 | Gear microscopic error vibration prediction method based on neural network |
Family Cites Families (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102506753B (en) * | 2011-11-04 | 2013-10-16 | 西安电子科技大学 | Fourteen-point spherical wavelet transformation-based shape difference detection method for irregular parts |
| CN103048135A (en) * | 2012-12-15 | 2013-04-17 | 新昌县冠阳技术开发有限公司 | Multi-fault coupling experimenter of flexible rotor rolling bearing foundation system, and fault identification method |
| CN103471841B (en) * | 2013-09-30 | 2015-08-12 | 国家电网公司 | A kind of rotating machinery vibrating failure diagnosis method |
| CN104155108B (en) * | 2014-07-21 | 2017-07-07 | 天津大学 | A kind of Fault Diagnosis of Roller Bearings based on vibration time frequency analysis |
| CN105225243B (en) * | 2015-10-15 | 2017-10-10 | 徐德明 | One kind can antimierophonic method for detecting image edge |
| CN106250812B (en) * | 2016-07-15 | 2019-08-20 | 汤一平 | A kind of model recognizing method based on quick R-CNN deep neural network |
| CN106295139B (en) * | 2016-07-29 | 2019-04-02 | 汤一平 | A kind of tongue body autodiagnosis health cloud service system based on depth convolutional neural networks |
| CN106384076A (en) * | 2016-08-29 | 2017-02-08 | 湖南文理学院 | Sobel operator Wigner-Hough transform based gear fault feature extraction method |
| CN106503642B (en) * | 2016-10-18 | 2019-09-20 | 长园长通新材料股份有限公司 | A kind of model of vibration method for building up applied to optical fiber sensing system |
| CN106600581B (en) * | 2016-12-02 | 2019-09-24 | 北京航空航天大学 | A kind of train operation automatic fault detection system and method based on binocular stereo vision |
| CN107784122A (en) * | 2017-11-22 | 2018-03-09 | 殷周平 | A kind of instance-level image search method represented based on multilayer feature |
-
2018
- 2018-06-07 CN CN201810576763.9A patent/CN108871760B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108871760A (en) | 2018-11-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108334936B (en) | Fault prediction method based on migration convolutional neural network | |
| CN109297689B (en) | Large-scale hydraulic machinery intelligent diagnosis method introducing weight factors | |
| CN106124212A (en) | Based on sparse coding device and the Fault Diagnosis of Roller Bearings of support vector machine | |
| CN108871760B (en) | Efficient gear fault mode identification method | |
| CN111783531B (en) | Water turbine set fault diagnosis method based on SDAE-IELM | |
| CN104202765A (en) | Method and device for detecting sensor faults based on distributed network | |
| CN113505655A (en) | Bearing fault intelligent diagnosis method for digital twin system | |
| US20210096530A1 (en) | System and method for identifying manufacturing defects | |
| CN111931601A (en) | System and method for correcting error class label of gear box | |
| CN115791174B (en) | Rolling bearing abnormality diagnosis method, system, electronic equipment and storage medium | |
| CN114266197A (en) | Method for diagnosing equipment fault of hydropower station | |
| KR102404498B1 (en) | Industrial gearbox failure diagnosis apparatus and method using convolutional neural network based on adaptive time-frequency representation | |
| CN114548199A (en) | Multi-sensor data fusion method based on deep migration network | |
| CN116012681A (en) | Method and system for diagnosing motor faults of pipeline robot based on sound vibration signal fusion | |
| CN116842379A (en) | Mechanical bearing residual service life prediction method based on DRSN-CS and BiGRU+MLP models | |
| CN114330430A (en) | Elevator fault judgment method and system based on big data characteristic analysis | |
| CN111982514B (en) | Bearing fault diagnosis method based on semi-supervised deep belief network | |
| CN113627268A (en) | Model training method, and method and device for detecting fault of speed reducer for mine hoist | |
| CN111539931A (en) | Appearance abnormity detection method based on convolutional neural network and boundary limit optimization | |
| CN110727669B (en) | Electric power system sensor data cleaning device and cleaning method | |
| CN108548669B (en) | Fault diagnosis method and system for transmission system of engineering equipment | |
| CN116400168A (en) | Power grid fault diagnosis method and system based on depth feature clustering | |
| CN115618263A (en) | Feature-level SMOTE-based complex equipment unbalance fault diagnosis method | |
| CN114841196A (en) | Mechanical equipment intelligent fault detection method and system based on supervised learning | |
| CN113158769A (en) | CEEMDAN and FastICA-based electromechanical device bearing vibration signal denoising method |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20211203 Address after: Industrial processing zone, Maonan Development Zone, Maoming City, Guangdong Province (courtyard, 239 maoshui Road) Patentee after: Guangdong Dawei Automobile Industry Co.,Ltd. Address before: 525000 No. two, No. 139, Guandu Road, Guangdong, Maoming Patentee before: GUANGDONG University OF PETROCHEMICAL TECHNOLOGY |
|
| TR01 | Transfer of patent right |