CN103994062A - Hydraulic-pump fault feature signal extraction method - Google Patents
Hydraulic-pump fault feature signal extraction method Download PDFInfo
- Publication number
- CN103994062A CN103994062A CN201410199895.6A CN201410199895A CN103994062A CN 103994062 A CN103994062 A CN 103994062A CN 201410199895 A CN201410199895 A CN 201410199895A CN 103994062 A CN103994062 A CN 103994062A
- Authority
- CN
- China
- Prior art keywords
- signal
- hydraulic pump
- pump fault
- hydraulic
- component
- 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
Abstract
The invention provides a hydraulic-pump fault feature signal extraction method based on local main value decomposition. Endpoint continuation processing is conducted on collected hydraulic-pump fault signals with a related image method, so that signal endpoints are extreme points; smooth estimation is conducted on extreme values and local mean values with the method of fitting an envelope line through a cubic B spline, so that fitting errors are reduced; then, LMD decomposition is conducted on the preprocessed signals to obtain PF components, correlation analysis is conducted on PFs and the original signals, the real PF components are selected, and finally Hilbert transformation is conducted on the real PF components to obtain Hilbert spectrums of the signals, wherein the Hilbert spectrums are signal features of the hydraulic-pump fault signals. The method is suitable for fault diagnosis of a hydraulic pump in the industrial field.
Description
Technical field
The application provides a kind of Hydraulic pump fault characteristic signal extracting method of decomposing based on local mean value, is applicable to the fault diagnosis of mechanical field oil hydraulic pump.
Background technique
Oil hydraulic pump is the critical component in hydraulic system, and the whether good of its operation has important impact to whole hydraulic system, once oil hydraulic pump breaks down, can produce the phenomenons such as vibration, noise, performance is affected, and Efficiency Decreasing even can cause serious accident.Therefore, the fault characteristic signals of oil hydraulic pump is extracted in commercial Application and has great significance.
At present, mainly adopt the method for Digital Signal Analysis and Processing to realize the detection and diagnosis of oil hydraulic pump both at home and abroad.Oil hydraulic pump is the one of rotating machinery, and it can produce vibration in running, and vibration is periodic.If oil hydraulic pump breaks down, oscillating signal there will be periodic pulse so, and the frequency of this pulse is exactly failure frequency.Therefore, gather the oscillating signal of oil hydraulic pump, can extract signal characteristic by suitable method, thereby indirectly obtain the operation conditions of oil hydraulic pump.
At present, conventional feature extracting method has wavelet analysis method, empirical mode decomposition (EMD) method, local mean value to decompose (LMD) method etc.Wavelet analysis method is a kind of Time-Frequency Analysis Method, both can carry out global analysis to signal, can realize again partial analysis, but the selection of this and wavelet basis has much relations, once wavelet basis is selected, in analytic process, cannot change, the resolution of signal analysis is also just fixing so, and this explanation wavelet analysis does not have adaptability to the partial analysis of signal.EMD method is a kind of adaptive signal analysis method, and it is by several intrinsic mode functions of signal decomposition (IMF) sum, and each IMF is the signal of approximate unifrequency composition, and like this, the Hilbert that combines each IMF composes the frequency spectrum that can obtain whole signal.The problem that the method exists is, when IMF is carried out to Hilbert conversion, can produce unaccountable negative frequency, and can cause end effect in signal decomposition process, is mixed with the false component irrelevant with original signal in the IMF decompositing.
In sum, analysis of vibration signal is the effective method that oil hydraulic pump is carried out to detection and diagnosis.At present conventional wavelet method cannot self adaption adjustment in resolution, and EMD method exists the problems such as negative frequency, end effect and false component, need to study the adaptive signal analysis method of the existing problem of a kind of EMD of overcoming method.
Summary of the invention
In order to solve existing problem in said method, the application provides a kind of Hydraulic pump fault signal characteristic extracting methods that decomposes (LMD) based on local mean value, comprises four modules of layout, data capture, signal analysis and processing, feature extraction of sensor.By relevant image method, the Hydraulic pump fault signal gathering is carried out to end points continuation processing, signal is carried out to matching envelope by cubic B-spline method, then carry out LMD decomposition and obtain PF component, PF and original signal are carried out to correlation analysis, select real PF component, finally PF is carried out to Hilbert conversion, obtain the Hilbert spectrum of signal, be the signal characteristic of Hydraulic pump fault signal.
LMD decomposition is to grow up on the basis of decomposing at EMD, it is a series of PF (production function) component sums with physical significance by signal decomposition, PF component is multiplied each other and is obtained by envelope signal and pure FM signal, the instantaneous frequency being calculated by pure FM signal be positive, continuous, there is physical significance.Thereby solve the problem that occurs negative frequency when the IMF that EMD is decomposed carries out Hilbert conversion, in LMD, also existed the problem such as end effect, false PF component, adopted respectively self correlation image method, cross-correlation analysis screening method to process.
The concrete steps of Hydraulic pump fault signal characteristic extracting methods are as follows:
(1) rational placement sensor, the oscillating signal while gathering oil hydraulic pump operation;
(2) adopt self correlation image method to carry out continuation to signal two ends;
1) establish signal
length is n, and the corresponding time is followed successively by
left end point is
, and
not extreme point,
extreme point be followed successively by from left to right
, the corresponding time is followed successively by
, mistake
a signature waveform of waveform composition of these four points, is designated as
, length is L.It is at least comprising a maximum value minimum and Zero Crossing Point;
2) will
carry out autocorrelation calculation, formula is as follows:
In formula, N is signal length, and n is time delay number;
3) get
mirror image, continuation is to signal left end point, after continuation, signal left end point is an extreme point;
4) right endpoint is the same carries out continuation;
(3) before LMD decomposes, adopt cubic B-spline interpolation matching envelope, concrete steps are as follows:
Construct one and control polygonal, control polygonal and be formed by connecting by each control vertex, establish
for controlling the polygonal length of side, length of side total length is
.?
the knot vector of inferior B-spline curves can be written as:
In order to meet the local property requirement of B-spline curves, when calculating, control polygonal limit
corresponding by it
bar limit and replace, then to its standardization.Calculate definitional domain internal node siding-to-siding block length:
Taking given signal extreme point as control vertex
,
inferior B Splines Interpolation Curve control polygonal is corresponding
bar limit and
for:
The knot vector standardization denominator of k B Splines Interpolation Curve
for:
Can obtain k all nodal values of B spline-fit extremal by above analysis is:
The knot vector after parametrization can be expressed as:
Define and by known nodal value substitution equation matching Envelope equation be by B-spline Curve:
So just, try to achieve Envelope Equations;
(4) carry out LMD decomposition.Process is as follows:
1) find out signal two neighboring pole value point (comprising maximum and minimum)
, try to achieve local average by formula following formula respectively
with envelope estimated value
The local average and the envelope estimated value that adopt slipping smoothness to process gained are carried out smoothing processing, obtain respectively local mean value function
with envelope estimation function
;
2) by local mean value function
from primary signal, separate and obtain signal
, and right
carry out demodulation operation as follows:
3) if
do not meet
,
repeat as primary signal
inferior
step, until satisfy condition.Last local amplitude meets
, the first instantaneous amplitude that condition obtains while termination is designated as
,
First multiplicative function obtaining is:
4) by first
component is separated from primary signal,
Obtain
repeat as new signal
formula
inferior, until
for single signal.Last signal is broken down into
individual
component and a simple signal sum:
(5) adopt cross-correlation analysis screening method to reject false component.PF and original signal are carried out to cross-correlation analysis, and original signal and PF are done to normalization, suppose that the correlation coefficient obtaining is
, do following processing:
Setting threshold
,
be one and be greater than 1 real number, can select as the case may be; When
time, retain the
individual PF;
time, reject the
individual PF, and added to residual error part;
(6) PF that LMD is decomposed to gained does Hilbert analysis, is the signal characteristic of Hydraulic pump fault signal.
The application provide based on local mean value decomposed solution press pump trouble signal feature extracting method, by to end effect, envelope matching with eliminate the processing of false component, can extract fast and accurately the fault signature of oscillating signal, can carry out detection and diagnosis to oil hydraulic pump easily, for enterprise has saved a large amount of manpower and financial resources.
brief description of the drawings:fig. 1 is Hydraulic pump fault signal characteristic extracting methods flow chart; Fig. 2 is point position schematic diagram; Fig. 3 is LMD decomposition process figure.
Embodiment:
In accompanying drawing 2: 1. the axial measuring point of the vertical measuring point 3. of horizontal measuring point 2..
1. signals collecting: utilize acceleration transducer to carry out signals collecting to Hydraulic pump fault signal.In the application, sensor used is DH186IEPE piezoelectric acceleration transducer, and its sensitivity is 0 ~ 10mV/ms
-2, range is 500m/s
2, frequency range 0.5 ~ 5kHz.Vasculum is selected the DH-5923 dynamic signalling analysis instrument of eastern magnificent test company.Acceleration transducer is arranged in three directions, is respectively substantially horizontal, Vertical direction, axial.Be axially the centering degree in order to detect oil hydraulic pump, the signal of horizontal and vertical direction has reflected the fault signature of oil hydraulic pump.
2. adopt self correlation image method to carry out end points processing to signal;
If signal
length is n, and the corresponding time is followed successively by
left end point is
, and
not extreme point,
extreme point be followed successively by from left to right
, the corresponding time is followed successively by
, mistake
a signature waveform of waveform composition of these four points, is designated as
, length is L.It is at least comprising a maximum value minimum and Zero Crossing Point; Will
carry out autocorrelation calculation, formula is as follows:
In formula, N is signal length, and n is time delay number;
Get
mirror image, continuation is to signal left end point, after continuation, signal left end point is an extreme point; Right endpoint is the same carries out continuation.
3. adopt cubic B-spline interpolation matching envelope.In cubic B-spline interpolation, polygonal is formed by connecting by each control vertex, taking given signal extreme point as control vertex, distance between two control vertexs is the polygonal length of side, try to achieve all length of side sums of polygonal, drawn the knot vector of B-spline curves by formula, and then obtain all nodal values of B spline-fit extremal, and finally try to achieve the knot vector after parametrization, define and nodal value substitution Solving Equations is obtained to Envelope equation according to B-spline Curve;
Construct one and control polygonal, control polygonal and be formed by connecting by each control vertex, establish
for controlling the polygonal length of side, length of side total length is
.?
the knot vector of inferior B-spline curves can be written as:
In order to meet the local property requirement of B-spline curves, when calculating, control polygonal limit
corresponding by it
bar limit and replace, then to its standardization.Calculate definitional domain internal node siding-to-siding block length:
Taking given signal extreme point as control vertex
,
inferior B Splines Interpolation Curve control polygonal is corresponding
bar limit and
for:
The knot vector standardization denominator of k B Splines Interpolation Curve
for:
Can obtain k all nodal values of B spline-fit extremal by above analysis is:
The knot vector after parametrization can be expressed as:
Define and by known nodal value substitution equation matching Envelope equation be by B-spline Curve:
So just, try to achieve Envelope Equations.
4. pair treated signal carries out LMD decomposition;
1) find out signal two neighboring pole value point (comprising maximum and minimum)
, try to achieve local average by formula following formula respectively
with envelope estimated value
The local average and the envelope estimated value that adopt slipping smoothness to process gained are carried out smoothing processing, obtain respectively local mean value function
with envelope estimation function
;
2) by local mean value function
from primary signal, separate and obtain signal
, and right
carry out demodulation operation as follows:
3) if
do not meet
,
repeat as primary signal
inferior
step, until satisfy condition.Last local amplitude meets
, the first instantaneous amplitude that condition obtains while termination is designated as
,
First multiplicative function obtaining is:
4) by first
component is separated from primary signal,
Obtain
repeat as new signal
formula
inferior, until
for single signal.Last signal is broken down into
individual
component and a simple signal sum:
。
5. adopt cross-correlation analysis screening method to reject false component.PF and original signal are carried out to cross-correlation analysis, and original signal and PF are done to normalization, suppose that the correlation coefficient obtaining is
, do following processing:
Setting threshold
,
be one and be greater than 1 real number, can select as the case may be; When
time, retain the
individual PF;
time, reject the
individual PF, and added to residual error part.
6. the PF that LMD is decomposed to gained is Hilbert and analyzes, and is the signal characteristic of Hydraulic pump fault signal.
Claims (4)
1. a Hydraulic pump fault signal characteristic extracting methods, it is characterized in that, by relevant image method, the Hydraulic pump fault signal gathering is carried out to end points continuation processing, signal is carried out to matching envelope by cubic B-spline method, then carry out LMD decomposition and obtain PF component, PF and original signal are carried out to correlation analysis, select real PF component, finally real PF is carried out to Hilbert conversion, obtain the Hilbert spectrum of signal, be the signal characteristic of Hydraulic pump fault signal.
2. Hydraulic pump fault signal characteristic extracting methods as claimed in claim 1, is characterized in that the signal collecting to carry out end points continuation, adopts self correlation image method to process the signature waveform at two end points places, and realizing signal end is extreme point.
3. Hydraulic pump fault signal characteristic extracting methods as claimed in claim 1, is characterized in that when LMD decomposes adopting the method for B-spline curve envelope smoothly to estimate extreme value and local mean value, has reduced error of fitting.
4. Hydraulic pump fault signal characteristic extracting methods as claimed in claim 1, is characterized in that adopting cross-correlation analysis screening method to process PF and the original signal of LMD decomposition, rejects false PF component, obtains real PF component.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410199895.6A CN103994062A (en) | 2014-05-13 | 2014-05-13 | Hydraulic-pump fault feature signal extraction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410199895.6A CN103994062A (en) | 2014-05-13 | 2014-05-13 | Hydraulic-pump fault feature signal extraction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103994062A true CN103994062A (en) | 2014-08-20 |
Family
ID=51308343
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410199895.6A Pending CN103994062A (en) | 2014-05-13 | 2014-05-13 | Hydraulic-pump fault feature signal extraction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103994062A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104390781A (en) * | 2014-11-26 | 2015-03-04 | 中国矿业大学 | Gear fault diagnosis method based on LMD and BP neural network |
| CN104568024A (en) * | 2015-01-21 | 2015-04-29 | 山东理工大学 | Vibration type flow meter characteristic signal extraction method |
| CN104832418A (en) * | 2015-05-07 | 2015-08-12 | 北京航空航天大学 | Hydraulic pump fault diagnosis method based on local mean conversion and Softmax |
| CN105910805A (en) * | 2016-04-25 | 2016-08-31 | 电子科技大学 | Wavelet local mean decomposition method used for rotor rub-impacting fault diagnosis |
| CN106224224A (en) * | 2016-07-13 | 2016-12-14 | 北京航空航天大学 | A kind of based on Hilbert-Huang transform and quality the Hydraulic pump fault feature extracting method away from entropy |
| CN108664901A (en) * | 2018-04-20 | 2018-10-16 | 三峡大学 | Based on the micro-capacitance sensor power quality disturbance signal detection method for improving LMD |
| CN110907770A (en) * | 2019-11-28 | 2020-03-24 | 深圳供电局有限公司 | Partial discharge pulse feature extraction method and device, computer equipment and medium |
| CN114997242A (en) * | 2022-06-30 | 2022-09-02 | 吉林大学 | Extreme value positioning waveform continuation LMD signal decomposition method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040226346A1 (en) * | 2003-05-16 | 2004-11-18 | Eberhard Schluecker | Early fault detection in pump valves |
| CN102829974A (en) * | 2012-08-07 | 2012-12-19 | 北京交通大学 | LMD (local mean decomposition) and PCA (principal component analysis) based rolling bearing state identification method |
| CN102866027A (en) * | 2012-08-13 | 2013-01-09 | 燕山大学 | Rotary machinery fault feature extracting method based on local mean decomposition (LMD) and local time-frequency entropy |
| CN103291600A (en) * | 2013-03-18 | 2013-09-11 | 北京航空航天大学 | Fault diagnosis method for hydraulic pump based on EMD-AR (empirical mode decomposition-auto-regressive) and MTS (mahalanobis taguchi system) |
-
2014
- 2014-05-13 CN CN201410199895.6A patent/CN103994062A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040226346A1 (en) * | 2003-05-16 | 2004-11-18 | Eberhard Schluecker | Early fault detection in pump valves |
| CN102829974A (en) * | 2012-08-07 | 2012-12-19 | 北京交通大学 | LMD (local mean decomposition) and PCA (principal component analysis) based rolling bearing state identification method |
| CN102866027A (en) * | 2012-08-13 | 2013-01-09 | 燕山大学 | Rotary machinery fault feature extracting method based on local mean decomposition (LMD) and local time-frequency entropy |
| CN103291600A (en) * | 2013-03-18 | 2013-09-11 | 北京航空航天大学 | Fault diagnosis method for hydraulic pump based on EMD-AR (empirical mode decomposition-auto-regressive) and MTS (mahalanobis taguchi system) |
Non-Patent Citations (2)
| Title |
|---|
| 梁世杰等: "《局域均值分解中PF获取方法的改进》", 《航空计算技术》 * |
| 胡劲松等: "基于样条的振动信号局域均值分解方法", 《数据采集与处理》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104390781A (en) * | 2014-11-26 | 2015-03-04 | 中国矿业大学 | Gear fault diagnosis method based on LMD and BP neural network |
| CN104568024A (en) * | 2015-01-21 | 2015-04-29 | 山东理工大学 | Vibration type flow meter characteristic signal extraction method |
| CN104832418A (en) * | 2015-05-07 | 2015-08-12 | 北京航空航天大学 | Hydraulic pump fault diagnosis method based on local mean conversion and Softmax |
| CN105910805A (en) * | 2016-04-25 | 2016-08-31 | 电子科技大学 | Wavelet local mean decomposition method used for rotor rub-impacting fault diagnosis |
| CN105910805B (en) * | 2016-04-25 | 2018-06-01 | 电子科技大学 | A kind of small echo part mean decomposition method for Rotor Rubbing Fault diagnosis |
| CN106224224A (en) * | 2016-07-13 | 2016-12-14 | 北京航空航天大学 | A kind of based on Hilbert-Huang transform and quality the Hydraulic pump fault feature extracting method away from entropy |
| CN108664901A (en) * | 2018-04-20 | 2018-10-16 | 三峡大学 | Based on the micro-capacitance sensor power quality disturbance signal detection method for improving LMD |
| CN108664901B (en) * | 2018-04-20 | 2021-04-13 | 三峡大学 | Improved LMD-based micro-grid power quality disturbance signal detection method |
| CN110907770A (en) * | 2019-11-28 | 2020-03-24 | 深圳供电局有限公司 | Partial discharge pulse feature extraction method and device, computer equipment and medium |
| CN114997242A (en) * | 2022-06-30 | 2022-09-02 | 吉林大学 | Extreme value positioning waveform continuation LMD signal decomposition method |
| CN114997242B (en) * | 2022-06-30 | 2023-08-29 | 吉林大学 | Extremum positioning waveform extension LMD signal decomposition method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103994062A (en) | Hydraulic-pump fault feature signal extraction method | |
| CN102539150B (en) | Self-adaptive failure diagnosis method of rotary mechanical component based on continuous wavelet transformation | |
| CN107402131B (en) | A kind of bullet train kinematic parameter recognition methods based on rail side acoustic signal time-frequency crestal line | |
| CN105588717A (en) | Gearbox fault diagnosis method | |
| CN104133982A (en) | Elimination method of slurry pulse signal pump stroke noise | |
| CN104820786A (en) | Method for analyzing instantly weighted synchronous extrusion wavelet bispectrum | |
| DE112009005228T5 (en) | Ultrasonic detection device | |
| CN104597502A (en) | Novel petroleum seismic exploration data noise reduction method | |
| CN104217112A (en) | Multi-type signal-based power system low-frequency oscillation analysis method | |
| CN108416282B (en) | Method for extracting acoustic velocity of echo signal of underground working fluid level based on tubing coupling | |
| CN101655834A (en) | Signal separation method based on fractional wavelet transform | |
| CN111769810A (en) | Fluid mechanical modulation frequency extraction method based on energy kurtosis spectrum | |
| CN113901379A (en) | Dynamic online rapid processing method for real-time data of edge terminal | |
| CN105354594A (en) | Mixing matrix estimation method aiming at underdetermined blind source separation | |
| JP5354505B2 (en) | Signal detection device, signal detection method, and signal detection device manufacturing method | |
| CN104665875A (en) | Ultrasonic Doppler envelope and heart rate detection method | |
| CN103886868A (en) | Impact sound detection method and detection system | |
| CN104048165A (en) | Pipeline leakage diagnosing method | |
| CN106772032B (en) | Fault feature extraction method for hydroelectric generating set | |
| CN114781466A (en) | Fault diagnosis method and system based on harmonic fundamental frequency of rotary mechanical vibration signal | |
| Liu et al. | Research on weak signal detection for downhole acoustic telemetry system | |
| CN105467270B (en) | Single Terminal Traveling Wave Fault Location back wave identification algorithm based on frequency spectrum similarity evaluation | |
| CN106595835A (en) | Sound beacon signal extraction method and system based on dislocation superposition algorithm | |
| CN111623986A (en) | Signal feature extraction method and system based on synchronous compression transformation and time-frequency matching | |
| CN107817083A (en) | A kind of more vibration source fault signal of mechanical equipment separators and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20140820 |
|
| RJ01 | Rejection of invention patent application after publication |