CN109916625A - A kind of single channel gear-box multiple faults separation double-core microprocessing systems - Google Patents
A kind of single channel gear-box multiple faults separation double-core microprocessing systems Download PDFInfo
- Publication number
- CN109916625A CN109916625A CN201910232499.1A CN201910232499A CN109916625A CN 109916625 A CN109916625 A CN 109916625A CN 201910232499 A CN201910232499 A CN 201910232499A CN 109916625 A CN109916625 A CN 109916625A
- Authority
- CN
- China
- Prior art keywords
- signal
- reconstruction
- module
- separation
- gear
- 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.)
- Granted
Links
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The present invention relates to a kind of double-core microprocessing systems of single channel gear-box multiple faults separation, belong to Gear Box Fault Diagnosis Technology and signal processing technology field.The present invention is the double-core microprocessing systems of DSP and ARM, wherein DSP core includes Parameters for Phase Space Reconstruction estimation module, phase space reconfiguration module, reconstruction parameter adjustment module, reconstruction signal separation module, fourier transformation module, and ARM kernel includes time-frequency domain spectrogram drafting module;The present invention is extended using time delay Reconstruction Method, avoids reconstruction signal using experience as the human intervention of judgment criteria;It is reduced using Principal Component Analysis because of the influence of noise and redundancy to reconstruction signal in gear-box;Using limited support sample kernel function and FastICA blending algorithm, source signal probability density function is estimated, and then obtain the nonlinear function for meeting source signal statistical property, it is final to realize that signal efficiently separates.The present invention greatly reduces reconstruction parameter in the separation of gear-box multiple faults improperly to be influenced with nonlinear function selection is separated.
Description
Technical field
The present invention relates to Gear Box Fault Diagnosis Technologies and signal processing analysis technical field, are related to a kind of single channel gear
Case multiple faults separates double-core microprocessing systems.
Background technique
Gear-box is the important mechanism of mechanical equipment transmitting power, meanwhile and mechanical fault diagnosis in important grind
Study carefully one of object.Therefore, the research for carrying out Gear Box Fault Diagnosis Technology has important meaning to the safe operation for ensureing mechanical equipment
Justice.And the difficult point of fault detection is combined failure, the estimation for the quantity of combined failure is even more the difficulty of combined failure diagnosis
Point.In order to grasp the operating condition of gear-box, general acceleration sensor picks up its vibration signal.However sensor can be only installed at
Gear case external surface, it is too long and compound for multiple vibration signals which results in the vibration signal transmission path of sensor acquisition,
The redundancy vibration of all parts is contained, except gear-box itself is there are the influence that redundancy is vibrated, there is also from acquisition channel
The influence in various electronic noises and ambient enviroment big noise disturbance source, causes the source number of each channel actual acquisition past
Toward the quantity for being more than sensor, which increase the difficulty of diagnosis, so generally installing multiple sensor acquisition letters in the lab
Number.But possibly multiple sensors can not be installed for cost problem and environmental problem in Practical Project problem, so single
Channel blind source separate technology scheme is used to diagnose gear-box mixed fault just the research hotspot become in recent years.
Most of blind source signal separation methods are only applicable to positive definite or overdetermination situation, so this special for single channel
Owe condition, the thinking for proposing virtual multichannel will owe to be converted into positive definite and overdetermination situation surely using channel expansion technique, such as
Application publication number CN109029973 A, date of publication are that the patent document on December 18th, 2018 discloses a kind of realization single channel
The method that gear-box mixes fault diagnosis uses empirical mode decomposition (EMD) for signal and is mapped to multiple intrinsic mode functions
(IMF), IMF component is chosen using kurtosis criterion and Correlation Coefficient Criteria be reconstructed into multi channel signals with source single channel signal again,
Extend single channel for multichannel.However, EMD has the anti-aliasing effect of difference, and when source signal is not natural mode of vibration
It can fail.In order to improve it, single channel signal is decomposed using set empirical mode decomposition (EEMD), such as Authorization Notice No. CN
107192553 B, date of publication are that disclose a kind of gear-box based on blind source separating compound for the patent document on March 2nd, 2018
Method for diagnosing faults carries out EEMD signal decomposition to the vibration signal addition white noise after noise reduction, overcomes anti-saw effect difference
It is insufficient.But for the evaluation index of none standard of the method for such extension multi channel signals, still rely on sentencing for the mankind
It is disconnected, it needs according to the prior information in relation to source signal.In gear distress detection, the C-C method of phase space reconstruction technique can be with
Mixed source is mapped to hyperspace without human intervention.Such as 106513879 B of Authorization Notice No. CN, date of publication is
The patent document on January 15th, 2019 discloses a kind of spark discharge state recognition and detection method based on chaology, adopts
Reconstruction parameter is found out with C-C method and then the phase space of discharge condition is reconstructed, and avoids the extension of human intervention signal.
The signal theoretically reconstructed is equivalent to source signal, but on condition that correctly selects Embedded dimensions and delay time, inappropriate selection
Reconstruction parameter will lead to reconstructed error.
In addition, also needing to separate to the signal of extension reconstruct, FastICA is based on independent analysis (ICA) algorithm development
A kind of self-adapting multi-channel blind source separating fast algorithm, the relatively broad application in blind source separating.Such as Authorization Notice No.
CN103575523 B, date of publication are that the patent document on December 9th, 2015 discloses one kind based on FastICA- spectrum kurtosis-packet
The rotary machinery fault diagnosis method of network spectrum analysis adds surveyed multichannel using the FastICA method based on negentropy maximization
Speed signal carries out decoupling separation, so that the mixing vibration signal collected is efficiently separated and come.However for FastICA
The selection of nonlinear function can have wrong choice in algorithm, so that FastICA algorithm falls into local extreme so that separation
Degradation, so the selection for nonlinear function is most important, but highly difficult.
Therefore, the selection for effectively reducing reconstruction parameter is improperly influenced with correct selection nonlinear function to more in gear-box
Source fault reconstruction is most important.
Summary of the invention
In order to reduce the improper influence to reconstruction signal of signal reconstruction parameter selection, while avoiding FastICA because of mistake
It selects nonlinear function and separating effect is had an impact.The invention proposes a kind of the double of single channel gear-box multiple faults separation
Core microprocessing systems, the double-core microprocessing systems can be analyzed and processed single pass acceleration vibration signal, will mix
Fault-signal in signal is effectively separated.
In order to realize the purpose of above-mentioned technology, the technical scheme of the present invention is realized as follows:
A kind of double-core microprocessing systems of single channel gear-box multiple faults separation are made of DSP and ARM kernel, in the double-core
Include Parameters for Phase Space Reconstruction estimation module, phase space reconfiguration, reconstruction parameter adjustment module, reconstruction signal point in microprocessing systems
From module, fourier transformation module, time-frequency domain spectrogram drafting module.
The DSP and ARM of the double-core microprocessing systems are that the division of labor is specific to Data Analysis Services, are divided the work as follows:
DSP core receives external input signal, establishes processing data object, if x=(1,2 ... N) indicate one long
Degree is the time series of N;Firstly, Parameters for Phase Space Reconstruction estimation module solves the estimated value of Parameters for Phase Space Reconstruction, this estimation
Algorithm need to only guarantee time window τ using a kind of complementary representative algorithm C-C method of Embedded dimensions and time delayw=
(m-1) * τ is constant, and wherein m is Embedded dimensions, and τ is lag time;Phase space reconfiguration module is according to the phase space estimated
Reconstruction parameter m and τ are carried out the extension of multichannel using time-delay method to single channel signal, obtain one group after phase space reconstruction
Matrix X=(N- (m-1) * τ -1) × m;Reconstruction parameter adjusts module to the signal reconstructed according to reconstruction parameter according to a preliminary estimate
The adjustment of parameter is reconstructed to it using principal component analytical method, reduces extension multichannel by finding the principal component in X
The influence of noise and redundancy in signal, then signal is mapped to new hyperspace;Reconstruction signal separation module is to adjusting
New multidimensional phase space signal after whole believes multichannel using the FastICA blending algorithm based on limited support sample kernel function
It number is separated, obtains the time-frequency domain signal of multi-source failure;Fourier transformation module carries out Fourier to isolated time-domain signal
Corresponding frequency-region signal is arrived in transformation.
The time frequency signal for the separation signal that verification DSP final process obtains in ARM carries out the drafting of time-frequency, then with periphery
Communication equipment docked.
Error existing for selection for the parameter of State Space Reconstruction, up to the present without a kind of phase space reconfiguration
Parameter estimation algorithm can correctly estimate reconstruction parameter, so according to the algorithm for estimating of existing Parameters for Phase Space Reconstruction
For the phase space that the reconstruction parameter obtained is reconstructed there are reconstructed error, most important reconstructed error is exactly to wrap in reconstruction signal
Noise and redundancy are contained.And principal component analytical method can pass through searching using reconstruct multidimensional signal as original signal matrix
Reconstruction signal is mapped to new hyperspace by the main component of signal, reduces noise and redundancy.
To the concrete processing procedure of the adjustment module of Parameters for Phase Space Reconstruction in the DSP core of the double-core microprocessing systems
Include:
S41: the covariance matrix Cov (X) of reconstruction signal, covariance matrix Cov (X) such as formula of phase space matrix are solved
Shown in lower
Cov (x in Cov (X), in diagonal lineii) be each dimension itself variance.Cov(xij) (i ≠ j) be i dimension
With the covariance of j dimension.If τ is greater than its optimal value, noise is included in the phase space of reconstruct, i.e. Cov (xii) will be smaller.
If m is greater than its optimum value, it is too big to reconstruct size, and Cov (xij) can be larger.If Cov (xii) larger and Cov
(xij) (i ≠ j) be less than setting value, then terminate, no person is switched to lower step;
S42: eigenvalue λ ∈ R is calculatedm, feature vector P ∈ Rm×m;
S43: removal characteristic value only accounts for 15% characteristic value of characteristic value summation, i.e. removal noise and existence of redundant, leaves
The trouble unit of gear-box, final new feature vector will become P ' ∈ Rm×e(e < m);
S44: new reconstruction signal y (y ∈ R is regained according to new eigen vector((N-(m-1)τ-1)×e)), from
And by eliminating the smallest characteristic value and corresponding feature vector, redundancy dimension is reduced, so as to disclose true real system
Feature.
In order to isolate source signal from mixed signal, need to know accurate probability density function G (y), for
FsatICA is to replace G (y) using specific nonlinear function, however, in the case where gear-box this source signals are unknown,
Just become the big influence factor most of mixed signal separation, in order to solve this problem, institute for the selection of nonlinear function
Reconstruction signal separation module is stated using the FastICA blending algorithm based on limited support sample kernel function, algorithm estimates weight
Structure signal probability density function obtains the nonlinear function for meeting source signal statistical property, using negentropy as objective function, then ties
The advantage of FastICA algorithm batch processing calculation method is closed, and then fast and accurately realizes the blind separation of mixed signal.
Include: to the concrete processing procedure of reconstruct signal separation module in the DSP core of the double-core microprocessing systems
S51: centralization processing is carried out to reconstruction signal y, its mean value is made to become 0;
S52: the correlation of whitening processing removal data is carried out to signal;
S53: selection needs the number of signals m ' estimated, and the number of iterations p ← 1 is arranged;
S54: selection initial weight vector (random) Wn;
S55: reconstruction signal is ranked up with non-decreasing sequence, obtains y={ y1, y2... yN, pass through
The probability density function of limited support sample kernel function is estimated, in formulaHave for M
Limit supports sample { yi, yi+1... yi+M, and M indicates the length of window, i.e.,By the boundary for meeting limited support sample
Condition and the primary condition of probability density function obtain lowest-order kernel function
In formula, μi=(yi+M-yi)/2 obtain G according to the condition of kernel functioni=15 μi -5/ 16, finally obtain the general of random y
The difference quotient of rate density p (y):
To estimate probability density function, nonlinear function g is sought, based on the principle of " separating one by one ", it is a excellent to solve n
The solution of change problem
Constraint condition are as follows:According to Kuhn Tucker condition and | | W | |=1 can pass through f
(W)=E { zG (WTZ) } value of+β W=0 is solved to optimal solution, and β is constant, is solved using newton iteration formula;
S56: it enablesThe single order that wherein g ' () represents g () is led
Number;
S57: Newton iteration is reduced to by a series of simplified process Wn+1←Wn+1/||Wn+1||;
S58: if WnIt does not restrain, then return step S56, otherwise continues;
S59: enabling n=n+1, if n≤m ', return step S54.
The technical effects of the invention are that being carried out using each core intrinsic advantage in dual core using DSP core on hardware
Parameters for Phase Space Reconstruction estimation, phase space reconfiguration, reconstruction parameter adjusts, reconstruction signal separates, fourier transformation module, in ARM
Core carries out separation signal time-frequency domain spectrogram and draws.On software conventional one-channel blind source separating there are aiming at the problem that, propose
Parameter adjustment is carried out to the signal after reconstruct using principal component analytical method, to reduce the noise and redundancy in reconstruction signal
It is inaccurate and extensive finally to have selected nonlinear function using wider FastICA separation algorithm to tradition for the influence of signal
The source signal appeared again is undesirable, using the FastICA blending algorithm based on limited support sample kernel function, to solve this
It is effective finally can to provide a high speed for this environment of the gear-box diagnosis object more complicated with operation conditions for a problem
Processing system.
Detailed description of the invention
Fig. 1 is system module block diagram of the invention;
Fig. 2 is the flow chart that the present invention uses method;
Fig. 3 is the flow chart of reconstruction parameter adjustment module of the present invention;
Fig. 4 is the flow chart of reconstruction signal separation module of the present invention.
Specific embodiment
For the clearer description present invention, the present invention is described in further detail with reference to the accompanying drawing and says
It is bright.
Referring to Fig. 1, which is made of DSP and ARM kernel, includes phase in the double-core microprocessing systems
Space Reconstruction parameter estimation module, phase space reconfiguration, reconstruction parameter adjust module, reconstruction signal separation module, time-frequency domain spectrogram
Drafting module.Wherein Parameters for Phase Space Reconstruction estimation module, phase space reconfiguration, reconstruction parameter adjustment module, reconstruction signal separation
Module, fourier transformation module belong to DSP core operation module, and time-frequency domain spectrogram drafting module belongs to ARM module.
Referring to fig. 2, which receives external input signal, establishes place to the process flow diagram of signal, DSP core
Data object is managed, if the time series that x=(1,2 ... N) indicates that a length is N is process object.Firstly, in DSP core
Parameters for Phase Space Reconstruction estimation module according to the acceleration vibration signal of single pass gear-box solve Parameters for Phase Space Reconstruction
Estimated value, this algorithm for estimating only needed using a kind of complementary representative algorithm C-C method of Embedded dimensions and time delay
Guarantee time window τw=(m-1) * τ is constant, and wherein m is Embedded dimensions, and τ is lag time;Phase space reconfiguration module according to
The Parameters for Phase Space Reconstruction m and τ estimated carries out the extension of multichannel to single channel signal using time-delay method, reconstructs phase
One group of matrix X=(N- (m-1) * τ -1) × m is obtained behind space;
Referring to Fig. 3, reconstruction parameter adjusts the flow chart of module, and reconstruction parameter adjusts module to according to reconstruct according to a preliminary estimate
Parameter and the signal that reconstructs it are reconstructed using principal component analytical method the adjustment of parameter, solve the covariance of reconstruction signal
Shown under Matrix C ov (X), the covariance matrix Cov (X) of phase space matrix such as formula:
If noise is included in the phase space of reconstruct, i.e. Cov (x since τ is greater than its optimal valueii) will be smaller.Such as
When fruit m is greater than its optimum value, then it is too big to reconstruct size, and Cov (xij) can be larger, so checking Cov (xii) and Cov (xij)
The size of (i ≠ j), Cov (xii) larger and Cov (xij) (i ≠ j) be less than setting value, then terminate, no person is switched to lower step;Then
Calculate the eigenvalue λ ∈ R of the matrixm, feature vector P ∈ Rm×m;Removal characteristic value only accounts for 15% feature of characteristic value summation
Value, i.e. removal noise and existence of redundant leave gear-box suspected of the characteristic value of trouble unit, and final new feature vector will become
P ' ∈ Rm×e(e < m);New reconstruction signal y (y ∈ R is regained according to new eigen vector((N -(m-1)τ-1)×e)), so that redundancy dimension is reduced, to reconstruct ginseng by eliminating the smallest characteristic value and corresponding feature vector
Number adjustment module completes the adjustment of the reconstruction parameter to phase space reconstruction signal.
Above-mentioned process completes the extension from single channel to multichannel, and from owing shape conversion for positive definite or overdetermination feelings
The process of the migration fractionation signal of shape, then after operation be exactly that blind source separating is carried out to multi channel signals.
With reference to Fig. 4, the specific flow chart of reconstruction signal separation module, reconstruction signal separation module is to new after being adjusted
Multidimensional phase space signal use the FastICA blending algorithm based on limited support sample kernel function.
S51: centralization processing is carried out to reconstruction signal y, its mean value is made to become 0;
S52: the correlation of whitening processing removal data is carried out to signal;
S53: selection needs the number of signals m ' estimated, and the number of iterations p ← 1 is arranged;
S54: selection initial weight vector (random) WnService is provided for subsequent Newton iteration;
S55: pre-treatment signal y is ranked up with non-decreasing sequence, obtains y={ y1, y2... yN, pass through
The probability density function of limited support sample kernel function is estimated, in formulaHave for M
Limit supports sample { yi, yi+1... yi+M, and M indicates the length of window, i.e.,By the boundary for meeting limited support sample
Condition and the primary condition of probability density function obtain lowest-order kernel function
In formula, μi=(yi+M-yi)/2 obtain G according to the condition of kernel functioni=15 μi -5/ 16, finally obtain the general of random y
The difference quotient of rate density p (y):
To estimate probability density function, nonlinear function g is sought, based on the principle of " separating one by one ", it is a excellent to solve n
The solution of change problem isConstraint condition isAccording to Kuhn-Tucker condition and | | W | |=1 can pass through f (W)=E { zG (WTz)}+βW
=0 value is solved to optimal solution, and β is constant, is solved using newton iteration formula;
S56: it enablesThe single order that wherein g ' () represents g () is led
Number;
S57: Newton iteration is reduced to by a series of simplified process Wn+1←Wn+1/||Wn+1||;
S58: judge WnConvergence, if WnThen return step S56 is dissipated, otherwise continue to the next step;
S59: enabling n=n+1, if n≤m ', return step S54.
It is finally recovered the source signal that is out of order, N={ N1, N2…NL, L is estimation source of trouble number }.Fourier transformation module to point
Time-domain signal from after carries out Fourier transformation, and then obtains corresponding frequency-region signal.
The time frequency signal for the separation signal that verification DSP final process obtains in ARM carries out the drafting of time-frequency, then with periphery
Communication equipment docked.
Claims (3)
1. a kind of single channel gear-box multiple faults separates double-core microprocessing systems, which is characterized in that the microprocessing systems by DSP with
ARM dual core composition includes Parameters for Phase Space Reconstruction estimation module, phase space reconfiguration, reconstruct in the double-core microprocessing systems
Parameter adjustment module, reconstruction signal separation module, fourier transformation module, time-frequency domain spectrogram drafting module;
DSP core in the double-core processing system complete the input of signal, Parameters for Phase Space Reconstruction estimation, phase space reconfiguration,
Reconstruction parameter adjustment, reconstruction signal separation, Fourier transformation, the ARM kernel completes the drafting of time-frequency domain spectrogram and periphery connects
The communication work of mouth;
The phase space reconfiguration module, with time-delay method to signal reconstruction, if single channel signal x=(1,2 ... N) it is one
Length is the time series of N, obtains one group of matrix X=(N- (m-1) * τ -1) × m after phase space reconstruction, and wherein m is insertion dimension
Number, τ is lag time, uses C-C method for reconstruction parameter estimation module, need to guarantee time window τw=(m-1) * τ is constant i.e.
It can;
The reconstruction parameter adjusts module, is mainly adjusted by principal component analytical method to reconstruction signal X, passes through removal
Characteristic value accounts for 15% characteristic value and corresponding feature vector of characteristic value summation in X, to reduce in extension multi channel signals
Noise and redundancy influence, signal is mapped to according to new eigen vector by new hyperspace y ∈ R again((N -(m-1)τ-1)×e), wherein e < m;
The reconstruction signal separation module is believed using the FastICA blending algorithm based on limited support sample kernel function from reconstruct
It number sets out, estimates the probability density function of the limited support sample kernel function of reconstruction signal
Wherein, φ () be lowest-order kernel function, meet limited support sample boundary condition and probability density function it is basic
The lowest-order kernel function φ (y) of condition
In formula, μi=(yi+M-yi)/2 obtain G according to the condition of kernel functioni=15 μi -5/ 16, and then according to reconstruction signal probability
The difference quotient H (y) of density
Source signal probability density function is calculated, the nonlinear function based on source signal statistical property is finally obtained;
The gear-box multiple faults based on a kind of single channel blind source separation algorithm separates double-core microprocessing systems, processes
Journey is as follows:
S1: the vibration signal of single-channel multi-fault source gear-box is input to micro- double-core processing system;
S2: the Parameters for Phase Space Reconstruction estimation module based on C-C algorithm estimates analyzed single-channel multi-fault gear-box vibration
The reconstruction parameter m and τ of signal;
S3: the reconstruction parameter that the phase space reconfiguration module based on time-delay method is found out according to step S2, by the gear analyzed
Packing list channel signal is extended to multi channel signals X=(N- (m-1) * τ -1) × m;
S4: reconstruction parameter based on principal component analytical method adjusts module, lead to the reconstruction signal X for finding that step S3 is obtained it is main at
Point, reconstruction signal is mapped to new hyperspace the adjustment of parameter is reconstructed and exports multidimensional signal again;
S5: the reconstruction signal separation module based on limited support sample Kernels Yu FastICA blending algorithm estimates general
Rate density function obtains the nonlinear function based on reconstruction signal statistical property, using negentropy as objective function, in conjunction with
FastICA algorithm batch processing calculates, and the separation that reconstruction signal realizes signal is newly obtained to S4;
S6: fourier transformation module carries out Fourier transformation to the time-domain signal after separation, to obtain corresponding frequency spectrum;
S7: time-frequency domain spectrogram drafting module, which carries out time-frequency spectrum to time frequency signal obtained by S6, to be drawn, so according to time-frequency figure into
Row gearbox fault must be positioned and be quantified.
2. the double-core micro process of the gear-box multiple faults separation according to claim 1 based on single channel blind source separation algorithm
System, which is characterized in that the step S4 includes:
S41: the covariance matrix Cov (X) of reconstruction signal is solved, if Cov (xii) larger and Cov (xij) (i ≠ j) be less than setting
Value, then terminate, no person is switched to lower step;
S42: eigenvalue λ ∈ R is calculatedm, feature vector P ∈ Rm×m;
S43: removal characteristic value accounts for 15% characteristic value of characteristic value summation, and new feature vector has become P ' ∈ Rm×e, wherein e
< m;
S44: new reconstruction signal y, y ∈ R is regained according to new eigen vector((N-(m-1)τ-1)×e)。
3. the double-core micro process of the gear-box multiple faults separation according to claim 1 based on single channel blind source separation algorithm
System, which is characterized in that the step S5 includes:
S51: centralization processing is carried out to reconstruction signal y, its mean value is made to become 0;
S52: the correlation of whitening processing removal data is carried out to signal;
S53: selection needs the number of signals m ' estimated, and the number of iterations p ← 1 is arranged;
S54: random selection initial weight vector Wn;
S55: utilizing limited support sample Kernels, and estimated probability density function seeks nonlinear function g;
S56: it enablesWherein g ' () represents the first derivative of g ();
S57: according toWn+1←Wn+1/||Wn+1||;
S58: if WnIt does not restrain, then returns to S56, otherwise continue;
S59: enabling n=n+1, if n≤m ', return step S54.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910232499.1A CN109916625B (en) | 2019-03-26 | 2019-03-26 | Single-channel gear box multi-fault separation dual-core micro-processing system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910232499.1A CN109916625B (en) | 2019-03-26 | 2019-03-26 | Single-channel gear box multi-fault separation dual-core micro-processing system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109916625A true CN109916625A (en) | 2019-06-21 |
CN109916625B CN109916625B (en) | 2020-08-18 |
Family
ID=66966791
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910232499.1A Active CN109916625B (en) | 2019-03-26 | 2019-03-26 | Single-channel gear box multi-fault separation dual-core micro-processing system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109916625B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110378065A (en) * | 2019-07-29 | 2019-10-25 | 河海大学 | A kind of large deformation plate non-linear vibratory signal State Space Reconstruction based on normal parameter |
CN110608884A (en) * | 2019-08-08 | 2019-12-24 | 桂林电子科技大学 | Rolling bearing state diagnosis method based on self-attention neural network |
CN111999607A (en) * | 2020-07-20 | 2020-11-27 | 中国南方电网有限责任公司超高压输电公司广州局 | Method and device for separating partial discharge narrow-band interference blind source under single-channel signal |
CN114216676A (en) * | 2021-11-30 | 2022-03-22 | 上海海事大学 | Compound fault diagnosis method for planetary gearbox without tachometer under time-varying working condition |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20010061133A (en) * | 1999-12-28 | 2001-07-07 | 이계안 | A power transmit error estimating apparatus between final reduction gear and differential ring gear for ff cars |
CN103868694A (en) * | 2014-03-26 | 2014-06-18 | 东南大学 | Embedded variable-rotation-speed bearing fault diagnosis device |
CN104122086A (en) * | 2014-08-11 | 2014-10-29 | 长沙理工大学 | Kriging model based wind turbine gearbox fault diagnosis method |
CN108956131A (en) * | 2018-07-30 | 2018-12-07 | 西安英特迈思信息科技有限公司 | A kind of express locomotive EEF bogie gear-box intellectual monitoring unit |
CN109029973A (en) * | 2018-06-22 | 2018-12-18 | 石家庄铁道大学 | The method for realizing the diagnosis of single channel gear-box mixed fault |
-
2019
- 2019-03-26 CN CN201910232499.1A patent/CN109916625B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20010061133A (en) * | 1999-12-28 | 2001-07-07 | 이계안 | A power transmit error estimating apparatus between final reduction gear and differential ring gear for ff cars |
CN103868694A (en) * | 2014-03-26 | 2014-06-18 | 东南大学 | Embedded variable-rotation-speed bearing fault diagnosis device |
CN104122086A (en) * | 2014-08-11 | 2014-10-29 | 长沙理工大学 | Kriging model based wind turbine gearbox fault diagnosis method |
CN109029973A (en) * | 2018-06-22 | 2018-12-18 | 石家庄铁道大学 | The method for realizing the diagnosis of single channel gear-box mixed fault |
CN108956131A (en) * | 2018-07-30 | 2018-12-07 | 西安英特迈思信息科技有限公司 | A kind of express locomotive EEF bogie gear-box intellectual monitoring unit |
Non-Patent Citations (1)
Title |
---|
王向红 等: "基于小波包和独立分量分析的微弱多源故障声发射信号分离", 《上海交通大学学报》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110378065A (en) * | 2019-07-29 | 2019-10-25 | 河海大学 | A kind of large deformation plate non-linear vibratory signal State Space Reconstruction based on normal parameter |
CN110608884A (en) * | 2019-08-08 | 2019-12-24 | 桂林电子科技大学 | Rolling bearing state diagnosis method based on self-attention neural network |
CN111999607A (en) * | 2020-07-20 | 2020-11-27 | 中国南方电网有限责任公司超高压输电公司广州局 | Method and device for separating partial discharge narrow-band interference blind source under single-channel signal |
CN111999607B (en) * | 2020-07-20 | 2021-08-20 | 中国南方电网有限责任公司超高压输电公司广州局 | Method and device for separating partial discharge narrow-band interference blind source under single-channel signal |
CN114216676A (en) * | 2021-11-30 | 2022-03-22 | 上海海事大学 | Compound fault diagnosis method for planetary gearbox without tachometer under time-varying working condition |
Also Published As
Publication number | Publication date |
---|---|
CN109916625B (en) | 2020-08-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109916625A (en) | A kind of single channel gear-box multiple faults separation double-core microprocessing systems | |
EP3309782B1 (en) | Method, device and system for noise suppression | |
JP4464966B2 (en) | Virtual load monitoring system and method | |
CN105929331A (en) | Double-fed aerogenerator stator/rotor fault analysis diagnosis apparatus and method | |
CN109635318B (en) | Intelligent analysis redundancy design method for aero-engine sensor | |
CN105846857A (en) | Determining the signal quality of an electrical interconnect | |
CN112733612A (en) | Cross-domain rotating machinery fault diagnosis model establishing method and application thereof | |
CN110426191A (en) | A kind of method for diagnosing faults of anti-interference rotating machinery | |
CN115081331A (en) | Wind turbine generator running state abnormity detection method based on state parameter reconstruction error | |
WO2020130810A1 (en) | Method and monitoring system for generator slow-coherency online identification and dynamic tracking | |
CN111060844A (en) | Interturn short-circuit fault diagnosis method and device for high-speed train traction transmission system | |
Liu et al. | Alleviating cosmological tensions with a coupled scalar fields model | |
CN104408025A (en) | Over-determined blind signal separation method based on spectrum correction and device of over-determined blind signal separation method | |
CN111507305B (en) | Fractional order self-adaptive stochastic resonance bearing fault diagnosis method based on WCSNR | |
CN103089232A (en) | Methods for detecting logging equipment and correcting consistency | |
CN109765212B (en) | Method for eliminating asynchronous fading fluorescence in Raman spectrum | |
CN116561541A (en) | Random dynamic load identification model, training method of model and identification method | |
Agunloye et al. | Lag length specification in Engle-Granger cointegration test: a modified Koyck mean lag approach based on partial correlation | |
CN110070887B (en) | Voice feature reconstruction method and device | |
CN109029973A (en) | The method for realizing the diagnosis of single channel gear-box mixed fault | |
CN112616160B (en) | Intelligent short-wave frequency cross-frequency-band real-time prediction method and system | |
CN113311220B (en) | Voltage sag diagnosis method, system, medium and electronic equipment | |
CN115293090A (en) | Multi-harmonic source responsibility quantification method based on reconstruction data processing | |
CN113657208A (en) | Vibration signal underdetermined blind source separation method for solving unknown source number of wind power tower | |
CN113311364A (en) | Permanent magnet synchronous motor inverter open-circuit fault diagnosis method based on multi-core SVM |
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 |