CN111597682B - Method for predicting remaining life of bearing of gearbox of wind turbine - Google Patents
Method for predicting remaining life of bearing of gearbox of wind turbine Download PDFInfo
- Publication number
- CN111597682B CN111597682B CN202010290308.XA CN202010290308A CN111597682B CN 111597682 B CN111597682 B CN 111597682B CN 202010290308 A CN202010290308 A CN 202010290308A CN 111597682 B CN111597682 B CN 111597682B
- Authority
- CN
- China
- Prior art keywords
- failure
- bearing
- degradation
- model
- gearbox
- 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
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000006731 degradation reaction Methods 0.000 claims abstract description 59
- 230000015556 catabolic process Effects 0.000 claims abstract description 54
- 238000009826 distribution Methods 0.000 claims abstract description 20
- 230000002860 competitive effect Effects 0.000 claims abstract description 17
- 238000012544 monitoring process Methods 0.000 claims description 20
- 239000013598 vector Substances 0.000 claims description 16
- 238000000513 principal component analysis Methods 0.000 claims description 15
- 230000007850 degeneration Effects 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 8
- 230000005653 Brownian motion process Effects 0.000 claims description 7
- 238000007476 Maximum Likelihood Methods 0.000 claims description 6
- 238000009792 diffusion process Methods 0.000 claims description 5
- 238000010586 diagram Methods 0.000 claims description 4
- 238000000342 Monte Carlo simulation Methods 0.000 claims description 3
- 238000005537 brownian motion Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 abstract description 5
- 238000013210 evaluation model Methods 0.000 abstract description 2
- 230000004927 fusion Effects 0.000 abstract description 2
- 238000011156 evaluation Methods 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 238000012423 maintenance Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000007797 corrosion Effects 0.000 description 2
- 238000005260 corrosion Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000005461 lubrication Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 241000039077 Copula Species 0.000 description 1
- 238000004026 adhesive bonding Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000003412 degenerative effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000007373 indentation Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 239000010687 lubricating oil Substances 0.000 description 1
- 238000012847 principal component analysis method Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
- 230000009528 severe injury Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 238000010977 unit operation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- 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
- G01M13/04—Bearings
-
- 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
- G01M13/04—Bearings
- G01M13/045—Acoustic or vibration analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/10—Numerical modelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Acoustics & Sound (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention discloses a method for predicting the residual service life of a bearing of a gearbox of a wind turbine, which is a method for predicting the residual service life of the bearing of the gearbox based on a competitive improved proportional failure model. The method comprises the following steps: carrying out principal component fusion on temperature and vibration data of a gearbox bearing in a whole life cycle from use to failure into an dimensional characteristic quantity, and reflecting the degradation performance of the bearing; and (3) taking the particularity of a gearbox bearing of the wind turbine generator into consideration, and establishing a competitive failure reliability evaluation model of Wiener degradation process and Weibull distribution sudden failure by adopting an improved proportional hazard model. Through example analysis, residual life analysis of a competitive model of degradation failure and sudden failure, an irrelevant model and an independent degradation failure and sudden failure model is compared, and the reliability and optimality of the competitive model are verified. The invention provides an effective, practical and high-precision prediction method for analyzing the residual service life of the bearing of the gearbox of the wind turbine.
Description
Technical Field
The invention relates to a method for predicting the residual life of a bearing of a gearbox of a wind turbine.
Background
With the increasing of the installed capacity of megawatt wind turbine generators, the gear box becomes an indispensable part, and the gearbox has the advantages of effectively improving the speed increasing ratio of a transmission system, reducing the occupied space of a cabin and optimizing the overall design of the cabin. However, the damage rate of the gearbox of the chinese wind power plant is as high as over 50%, and about 80% of the mechanical failures of the gearbox are caused by bearing failures.
The traditional wind turbine gearbox bearing research method is to analyze the service life of a bearing by utilizing a finite element technology, along with the development of a sensing technology and a signal processing technology, the research on the bearing is focused on state monitoring evaluation and fault diagnosis prediction, and the fault diagnosis method comprises an envelope spectrum method, an artificial neural network, a support vector machine and the like; in the aspect of state monitoring and evaluation, methods such as an intelligent method based on a neural network, fuzzy comprehensive judgment, matter element analysis and the like are adopted. A general bearing fault detection technology based on vibration signals is mostly adopted for a gearbox bearing in a wind power plant, fault data are easy to identify and single in data, and early fault identification and service life prediction of the bearing cannot be embodied.
From historical data of bearing faults of the wind power plant, more than 90% of faults of the bearing occur on the inner ring and the outer ring, and other faults basically occur on the rolling body and the retainer. In the operation process of the wind power bearing, due to the mutual influence of random factors such as improper lubrication, misalignment of a rotating shaft, instant wind speed load impact, emergency brake impact and the like, not only the degradation failure but also the sudden failure occurs in the aspect of the failure mode of the bearing. The degradation failure modes mainly comprise wear failure, fatigue failure, indentation failure, corrosion failure and gluing failure, and the sudden failure modes comprise fracture failure and retainer breakage.
Prediction and Health Management (PHM) has been proven to be a feasible path in many fields of aerospace, civil aircraft, nuclear reactors, and automobiles, and visual maintenance of the system is facilitated by early warning of system failure of long-life and high-reliability components. The current studies describing modeling of the degeneration process fall into two categories. Firstly, whether independent competitive failure modeling is carried out between the degradation failure and the burst failure is considered, and the accuracy of product reliability evaluation is difficult to accurately describe by adopting a simple series model. Secondly, considering the research that the performance degradation of the product has influence on the sudden failure, the method can be divided into two conditions: (1) By adopting various models such as a proportional hazard model, a degradation threshold-impact model, a Bayesian linear model, a state space model and the like, a plurality of product performance parameters are fused into a performance parameter model which can describe the overall degradation state of the product and is limited to the condition that a linear relation directly exists between the product parameters and the degradation quantity; and (2) establishing a combined life distribution model of a plurality of parameters. The Copula function is mostly adopted for binary relevance degradation modeling, and multivariate normal distribution is adopted for describing the relevance among a plurality of performance parameters.
Current research only describes the wind turbine generator bearing in terms of degradation failure, and does not consider the factor of sudden failure. How to utilize wind-powered electricity generation field real time monitoring data to establish the remaining life prediction evaluation of wind turbine gearbox bearing, utilize the predictive analysis on statistics through the method of data drive, manage the uncertainty of wind turbine generator system bearing trouble more effectively, combine expert's knowledge and solidify it in software simultaneously, form sustainable iterative intelligent system and maintain looking at the situation, reduce unit operation and maintenance cost, be the problem that the wind-powered electricity generation field awaits a solution urgently.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the method for predicting the residual life of the bearing of the gearbox of the wind turbine can predict the life of the bearing more accurately.
In order to solve the technical problems, the technical scheme adopted by the invention is as follows: a method for predicting the residual service life of a gearbox bearing based on a competitive improved proportional failure model comprises the following steps:
s1, wind turbine gearbox bearing characteristic parameter selection based on data driving
An SCADA system equipped in a wind power plant is used for carrying out real-time online monitoring on a unit gearbox bearing, and temperature and vibration monitoring parameters are used as main characteristic quantities;
s2, principal component analysis of characteristic parameters
S21, selecting characteristic parameters
Data extracted from the SCADA system are subjected to data-driven principal component analysis, and bearing temperature and amplitude data are fused into one dimensional characteristic quantity to reflect the degradation performance of the bearing;
s22. Principal component analysis
Extracting time-series data (C) of bearing temperature and amplitude 0, A 0 )、(C 1 ,A 1 )、(C 2 ,A 2 )、…、(C m ,A m ) Centralizing the characteristic parameters at each time
S23.PCA projection
After centralization, the characteristic values and the corresponding characteristic vectors are solved by utilizing a covariance matrix, the characteristic vector with the maximum characteristic value is selected for projection, and a state curve schematic diagram after PCA projection is obtained;
s3, establishing an improved competitive proportion danger model:
the failure time of Wiener degradation failure is T d The failure time of Weibull distribution burst failure is T r The failure time T of a bearing can be expressed as:
T＝min{T d ，T r }
the reliability of the failure of the bearing during the running time (0, t) can be expressed as R (t)
R(t)＝P{T>t}＝P{T d >t,T r >t}
Selecting a proportional failure model to establish a failure relation, and assuming that the bearing is not failed at the moment t, adopting an improved competitive proportional hazard model to establish the following model:
λ i (t,x t )＝h r (t)exp(β 0 α+β 1 δ i x t )
in the formula, h r (t) is a reference time burst failure rate function; beta is a 0 Is an initial value of degradation, beta 1 Is an increased failure rate parameter, alpha is a random vector coefficient, delta i Is a regression coefficient vector;
when considering the competition of the degeneration failure and the burst failure, the failure rate is
A degree of reliability of
When the degenerate failure and the burst failure are not related, the reliability is
S4, calculating model parameters:
s41, estimating degradation failure parameters
According to the degraded track of the bearing, the random vector coefficient alpha of each degraded track of the bearing can be obtained i And the regression coefficient vector δ i Calculating α i Can be obtained as an average of
δ i Obeying a normal distribution whose expectation and variance are +>
And &>
Assuming that M bearings are degraded and failed, at monitoring time t ij (i＝1,2，…，M；j＝1,2，…，M i ) Is marked as X (t) ij ) Establishing a likelihood function of the bearing degradation degree data according to the probability density function PDF of the Wiener process
Wherein σ is a diffusion coefficient, and μ is a drift coefficient; taking the logarithm of both sides, let mu and sigma 2 Are all equal to 0, resulting in μ and σ 2 Maximum likelihood estimation of
S42, burst parameter estimation
If the degradation failure and the burst failure are independent, estimating parameters according to the probability density function PDF of respective distribution;
if the degradation failure and the burst failure compete with each other, assuming that N bearings have burst failure, at the monitoring time t i The degree of bearing degradation of (i =1,2, \8230;, N) is noted as X (t) i ) Establishing the following likelihood function according to the probability density function PDF of the competition process
Taking the logarithm of two sides of the formula, wherein n is a shape parameter; η scale parameter; let beta 0 、β 1 Partial derivatives of n and eta are all equal to 0, and the maximum likelihood estimation of the parameters is realized by a Monte Carlo simulation method.
As a preferred scheme, the building of the Wiener degradation failure model specifically comprises:
if the failure threshold value of the bearing of the gearbox of the wind turbine generator is l, the time T of l is reached for the first time D -inf { t | X (t) ≧ l }; probability density function of degradation failure of
X(t)＝X(0)+μt+σB(t)
Wherein μ is a drift coefficient; σ (σ > 0) is a diffusion coefficient; b (-) is standard Brownian motion; x (0) is an initial value, the probability density function PDF of X (t) is
Time to failure of degeneration T d Obeying Inverse-Gaussian distribution, T d Has a reliability function of
As a preferred scheme, the building of the Weibull distribution burst failure model specifically includes:
the sudden failure process of the unit component obeys the Weibull distribution of two parameters, and then the sudden failure rate function h r (t) is
In the figure, n is a shape parameter; η scale parameter;
the probability density function PDF of h (t) is
Time to burst failure T r Has a reliability function of
The invention has the beneficial effects that:
1. the reliability evaluation method for the bearing of the gearbox of the wind turbine generator is provided, wherein the reliability evaluation method considers the existence of competitive failure between degradation failure and sudden failure;
2. combining with wind power plant SCADA real-time monitoring data, fusing bearing monitoring temperature and vibration data through a principal component analysis method to describe the degradation degree of a bearing;
3. and (3) considering the particularity of a gearbox bearing of the wind turbine generator, and establishing a competitive failure reliability evaluation model of a Wiener degradation process and Weibull distribution sudden failure by adopting an improved proportional hazard model.
Through the points, compared with the prior art, the method has the advantages that the residual service life prediction effect on the bearing of the gearbox of the wind turbine generator is better, and the precision is higher.
Drawings
FIG. 1 is an overall flow chart of the present invention.
FIG. 2 is a schematic representation of the bearing condition curve after PCA projection.
Fig. 3 is a probability chart of the bearing failure prediction of 8 machine set gearbox based on the method provided by the invention in the embodiment.
Detailed Description
Specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
The invention provides a method for predicting the residual life of a bearing of a gearbox of a wind turbine, which has the specific flow chart shown in the attached figure 1 and comprises the following steps:
s1, wind turbine gearbox bearing characteristic parameter selection based on data driving
The SCADA system equipped in the wind power plant monitors the bearings of the gearbox of the unit in real time on line, and the state monitoring parameters of the bearings can be regarded as internal covariates causing the performance degradation of the bearings. According to the wind field monitoring data, after the bearing of the gearbox of the wind turbine generator system breaks down, the temperature is increased sharply, the vibration amplitude is increased, and the wind speed and the lubricating oil parameters are not changed greatly, so that the temperature and the vibration monitoring parameters are used as main characteristic quantities.
S2, principal component analysis of characteristic parameters
S21, selecting characteristic parameters
The data volume extracted from the SCADA system is large and shows volatility and randomness, under the condition that the temperature and vibration sensors generate two-dimensional time sequence data at the same time and at the same wind speed, in order to conveniently show the variation trend of the bearing, the temperature and vibration data of the bearing are fused into one-dimensional characteristic quantity by adopting data-driven Principal Component Analysis (PCA), so that the degradation performance of the bearing is shown.
S23. Principal component analysis
Extracting time-series data (C) of bearing temperature and amplitude 0 ,A 0 )、(C 1 ,A 1 )、(C 2 ,A 2 )、…、(C m ,A m ) Centralizing the characteristic parameters at each time
S23.PCA projection
After the centering, the feature value and the corresponding feature vector are obtained by using the covariance matrix, and the feature vector with the largest feature value is selected for projection to obtain a state curve schematic diagram after PCA projection, as shown in FIG. 2.
S3, establishing an improved competitive proportion danger model:
s31.Wiener degeneration failure model
In consideration of the special situation of the geographical position of the wind turbine generator, the difference between the environmental temperature and the wind speed condition in four seasons is large, the bearing is influenced by low temperature, high temperature, corrosion, salt fog and other severe environmental factors, the overall degradation process of the bearing serving as a mechanical part rises along with the lapse of time, and the random performance degradation process can be well described by adopting a Wiener process.
If the failure threshold value of the bearing of the gearbox of the wind turbine generator is l, the time T of l is reached for the first time D ＝inf{t|X(t)≥l}。
Probability density function of degradation failure of
X(t)＝X(0)+μt+σB(t) (2)
Wherein μ is a drift coefficient; σ (σ > 0) is a diffusion coefficient; b (-) is standard Brownian motion; when X (0) is the initial value, the Probability Density Function (PDF) of X (t) is
Time to failure of degeneration T d Obeying Inverse-Gaussian distribution, T d Has a reliability function of
S32.Weibull distribution sudden failure model:
the failure occurs suddenly, and the normal operation of the whole gear box is further influenced. Historical fault statistics of the wind turbine generator shows that the sudden failure process of the wind turbine generator generally obeys two-parameter Weibull distribution, and then the sudden failure rate function h r (t) is
In the drawing, n is a shape parameter; and (6) eta scale parameters.
The PDF of h (t) is
Time to burst failure T r Has a reliability function of
S33. Competition proportion failure model
The failure time of the degenerative failure is T d The failure time of the burst failure is T r Regardless of the effects of lubrication and other factors, the failure time T of a bearing can be expressed as:
T＝min{T d ，T r } (8)
the reliability of the failure of the bearing during the running time (0, t) can be expressed as R (t)
R(t)＝P{T>t}＝P{T d >t,T r >t} (9)
The monitoring data of the wind turbine generator set comprises degradation data and service life data, so that a proportional failure model is selected to establish the establishment of a failure relation. Assuming that the bearing does not fail by the time t, considering the complexity of the bearing operation and the nonlinearity of the degradation process, an improved competitive proportion danger model is adopted [25] The following model is established:
λ i (t,x t )＝h r (t)exp(β 0 α+β 1 δ i x t ) (10)
in the formula, h r (t) is a reference time burst failure rate function; beta is a beta 0 Is an initial value of degradation, beta 1 Is an increased failure rate parameter, alpha is a random vector coefficient, delta i Is a regression coefficient vector.
When considering the competition of the degeneration failure and the burst failure, the failure rate is
A degree of reliability of
When the degenerate failure and the burst failure are uncorrelated, the reliability is
S4, model parameter calculation:
s41, estimating degradation failure parameters
According to the degradation track of the bearing, alpha of the degradation track of each bearing can be obtained i And delta i Calculating α i Can be obtained as an average of
δ i Obeying a normal distribution whose expectation and variance are +>
And &>
Assuming that there are M bearings undergoing degradation failure, at monitoring time t ij (i＝1,2，…，M；j＝1,2，…，M i ) Is marked as X (t) ij ) Establishing a likelihood function of the bearing degradation degree data according to the PDF of the Wiener process
Taking the logarithm of both sides, let mu and sigma 2 Are all equal to 0, resulting in μ and σ 2 Maximum likelihood estimation of
S42. Burst parameter estimation
If the degenerate failure and the bursty failure are independent of each other, the parameters can be estimated from the respective distributed PDFs.
If the degradation failure and the burst failure compete with each other, assuming that N bearings have burst failure, at the monitoring time t i The degree of bearing degradation of (i =1,2, \8230;, N) is noted as X (t) i ) From the PDFs of the competition process, the following likelihood functions are established
Taking the logarithm of both sides of formula (17) to make beta 0 、β 1 The partial derivatives of n and eta are all equal to 0, and the maximum likelihood estimation of the parameters is realized by a Monte Carlo simulation method.
In order to ensure that the calculation result is real and effective, 8 wind turbines in the same type of a certain wind farm in Banguchen, xinjiang are selected for research, the installation time of the 8 turbines is short, and the operation time and the environment are similar. The temperature and vibration data of the unit from self-running to fault can be directly acquired through SCADA on-line monitoring data. The cut-in wind speed of the unit is 3m/s, the cut-out wind speed is 22m/s, the temperature alarm threshold of a bearing of the gearbox is 80 ℃, the amplitude alarm threshold is 35mm, and SCADA system data are extracted every 4 hours. Data display in these units 7 # 、6 # 、3 # The bearing of the gear box of the machine set has sudden failure in 7.3, 7.5 and 7.6 years, the bearing of the gear box of the 8 machine sets has degradation failure in different degrees, the fusion state data of the bearing temperature and the amplitude is approximated to a linear regression model, the relevant parameters of the bearing of the gear box of each machine set are calculated, and the failure probability diagram of the bearing of the gear box of the 8 machine sets is calculated by considering a competition method, as shown in the attached figure 3.
For ease of analysis, the method proposed herein based on competition between degenerate failures and catastrophic failures is denoted as M 1 Method in which degeneration failure and burst failure are irrelevant is marked as M 2 The method considering only performance degradation failure is denoted as M 3 The method considering only burst failure is denoted as M 4 . Considering the wind farm 3 # 、6 # 、7 # All the 3 sets of gearbox bearings have sudden failures, and taking the 3 sets as an example, when the set runs for 26280h, the calculation is performed by the above four methods, and the predicted values of the residual life of the gearbox bearings are respectively compared, as shown in table 1.
TABLE 1 comparison of predicted results for three methods
As can be seen from Table 1, method M was used as the run time increased during actual operation 2 、M 3 And M 4 The error of the calculation gradually increases or decreases, method M 1 The calculation error of (2) is gradually decreasing. Method M 2 、M 3 And M 4 There is a risk of "underestimation" of bearing failure, which can lead to severe damage to gearbox bearings and further affect the operating conditions of the gearbox. Method M 2 The relevance of sudden failure and degradation failure is neglected, the degradation failure and the sudden failure of the bearing are simplified into a series model, and the prediction precision is low. Method M 3 Only the degradation condition of the bearing is considered, the particularity of the wind turbine generator is ignored, and the prediction precision is the lowest. Method M for considering only burst failure 4 The situation that the natural degradation of the bearing exists is not considered, and the situation that the prediction is inaccurate exists. Method M 1 Although the possibility of 'overestimating' the risk exists, the risk can be effectively controlled, and the probability of 'under-maintenance' of the bearing of the gearbox is improved. By comparison, the method provided by the invention has higher precision, comprehensively considers the fault risk and the maintenance cost, and is more reasonable and effective.
The above-mentioned embodiments are merely illustrative of the principles and effects of the present invention, and some embodiments may be used, not restrictive; it should be noted that various changes and modifications can be made by those skilled in the art without departing from the inventive concept, and these changes and modifications fall within the scope of the invention.
Claims (3)
1. A method for predicting the residual service life of a gearbox bearing based on a competitive improved proportional failure model comprises the following steps:
s1, wind turbine gearbox bearing characteristic parameter selection based on data driving
An SCADA system equipped in a wind power plant is used for carrying out real-time online monitoring on a unit gearbox bearing, and temperature and vibration monitoring parameters are used as main characteristic quantities;
s2, principal component analysis of characteristic parameters
S21, selecting characteristic parameters
Data extracted from the SCADA system are subjected to data-driven principal component analysis, and bearing temperature and amplitude data are fused into one dimensional characteristic quantity to reflect the degradation performance of the bearing;
s22. Principal component analysis
Extracting time-series data (C) of bearing temperature and amplitude 0 ,A 0 )、(C 1 ,A 1 )、(C 2 ,A 2 )、…、(C m ,A m ) Centralizing the characteristic parameters at each time
S23.PCA projection
After centralization, the characteristic values and the corresponding characteristic vectors are solved by utilizing a covariance matrix, the characteristic vector with the maximum characteristic value is selected for projection, and a state curve schematic diagram after PCA projection is obtained;
s3, establishing an improved competitive proportion danger model:
the failure time of the Wiener degradation failure model is T d The failure time of the Weibull distribution burst failure model is T r The failure time T of a bearing can be expressed as:
T＝min{T d ，T r }
the reliability of the failure of the bearing during the running time (0, t) can be expressed as R (t)
R(t)＝P{T>t}＝P{T d >t,T r >t}
Selecting a proportional failure model to establish a failure relation, and assuming that the bearing is not failed at the moment t, adopting an improved competitive proportional hazard model to establish the following model:
λ i (t,x t )＝h r (t)exp(β 0 α+β 1 δ i x t )
in the formula, h r (t) is a reference time burst failure rate function; beta is a 0 Is an initial value of degradation, beta 1 Is an increased failure rate parameter, α is a random vector coefficient, δ i Is a regression coefficient vector;
when considering the competition of the degeneration failure and the burst failure, the failure rate is
A degree of reliability of
When the degenerate failure and the burst failure are not related, the reliability is
S4, calculating model parameters:
s41, estimating degradation failure parameters
According to the degraded track of the bearing, the random vector coefficient alpha of each degraded track of the bearing can be obtained i And the regression coefficient vector δ i Calculating α i Can be obtained as an average of
δ i Obeying a normal distribution whose expectation and variance are +>
And &>
Assuming that there are M bearings undergoing degradation failure, at monitoring time t ij (i＝1,2，…，M；j＝1,2，…，M i ) Is marked as X (t) ij ) Establishing a likelihood function of the bearing degradation degree data according to the probability density function PDF of the Wiener process
Wherein, σ is a diffusion coefficient, and μ is a drift coefficient; taking the logarithm of both sides, let mu and sigma 2 Are all equal to 0, resulting in μ and σ 2 Maximum likelihood estimation of
S42, burst parameter estimation
If the degradation failure and the burst failure are independent, estimating parameters according to the probability density function PDF of respective distribution;
if the degradation failure and the burst failure compete with each other, assuming that N bearings have burst failure, at the monitoring time t i The degree of bearing degradation of (i =1,2, \8230;, N) is noted as X (t) i ) According to the probability density function PDF of the competition process, the following likelihood functions are established
Taking the logarithm of two sides of the formula, wherein n is a shape parameter; η scale parameter; let beta 0 、β 1 The partial derivatives of n and eta are all equal to 0, the maximum likelihood estimation of the parameters is carried out, and the correlation solving method is realized by a Monte Carlo simulation method.
2. The method for predicting the remaining service life of the gearbox bearing based on the competitive improved proportional failure model as claimed in claim 1, wherein the method comprises the following steps: the establishment of the Wiener degradation failure model specifically comprises the following steps:
if the failure threshold value of the bearing of the gearbox of the wind turbine generator is l, the time T of l is reached for the first time D -inf { t | X (t) ≧ l }; probability density function of degradation failure of
X(t)＝X(0)+μt+σB(t)
Wherein μ is a drift coefficient; σ (σ > 0) is a diffusion coefficient; b (-) is standard Brownian motion; x (0) is an initial value, the probability density function PDF of X (t) is
Time to failure of degradation T d Obeying an Inverse-Gaussian distribution, T d Has a reliability function of
3. The method for predicting the remaining service life of the gearbox bearing based on the competitive improved proportional failure model as claimed in claim 1, wherein the method comprises the following steps: the building of the Weibull distribution burst failure model specifically comprises the following steps:
the sudden failure process of the unit component obeys the Weibull distribution of two parameters, and then the sudden failure rate function h r (t) is
In the figure, n is a shape parameter; η scale parameter;
the probability density function PDF of h (t) is
Time to burst failure T r Has a reliability function of
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010290308.XA CN111597682B (en) | 2020-04-14 | 2020-04-14 | Method for predicting remaining life of bearing of gearbox of wind turbine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010290308.XA CN111597682B (en) | 2020-04-14 | 2020-04-14 | Method for predicting remaining life of bearing of gearbox of wind turbine |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111597682A CN111597682A (en) | 2020-08-28 |
| CN111597682B true CN111597682B (en) | 2023-03-31 |
Family
ID=72183163
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010290308.XA Active CN111597682B (en) | 2020-04-14 | 2020-04-14 | Method for predicting remaining life of bearing of gearbox of wind turbine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111597682B (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112417612B (en) * | 2020-10-15 | 2023-06-09 | 浙江工业大学 | Wind power gear box degradation state tracking and failure aggregation risk assessment method |
| CN114576102B (en) * | 2020-12-02 | 2024-09-20 | 金风科技股份有限公司 | Wind turbine generator system bearing state prediction method, device, equipment and storage medium |
| CN112632773B (en) * | 2020-12-21 | 2024-04-05 | 北京华能新锐控制技术有限公司 | Reliability prediction method for wind turbine generator |
| KR102525703B1 (en) * | 2021-01-26 | 2023-04-26 | 서울교통공사 | Apparatus for predicting failure of electric railway vehicle and method thereof |
| CN113095693B (en) * | 2021-04-19 | 2024-02-06 | 重庆理工大学 | Wind power bearing health state assessment method based on temperature characteristic quantity |
| US12112588B2 (en) * | 2021-04-28 | 2024-10-08 | Dana Automotive Systems Group, Llc | Systems and methods for prediction of component degradation |
| CN113297967B (en) * | 2021-05-24 | 2023-05-23 | 扬州大学 | Built-in reed harvester safety assessment method based on characteristic confidence range |
| CN113449985B (en) * | 2021-06-25 | 2023-04-07 | 上海电气风电集团股份有限公司 | Gearbox monitoring method and system and computer readable storage medium |
| CN113607580B (en) * | 2021-08-10 | 2023-09-05 | 江苏徐工工程机械研究院有限公司 | Fatigue test method and residual life prediction method for metal component |
| CN113819958A (en) * | 2021-11-22 | 2021-12-21 | 三一重型装备有限公司 | Predictive maintenance system, predictive maintenance method and heading machine |
| CN114326475B (en) * | 2021-11-30 | 2022-11-08 | 安徽海博智能科技有限责任公司 | Unmanned area controller reliability optimization method based on competitive failure mode |
| CN114676602B (en) * | 2022-05-27 | 2022-08-23 | 浙江大学 | Improved motor insulation multi-stress reliability assessment method and system |
| CN115935622A (en) * | 2022-11-25 | 2023-04-07 | 武汉科技大学 | Axial plunger pump residual life prediction method based on joint model identification |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5679883A (en) * | 1992-10-19 | 1997-10-21 | Wedeven; Lavern D. | Method and apparatus for comprehensive evaluation of tribological materials |
| WO2001084315A2 (en) * | 2000-05-04 | 2001-11-08 | The Procter & Gamble Company | Apparatus and process for simulating and analyzing a system with failure modes |
| WO2001084254A2 (en) * | 2000-05-04 | 2001-11-08 | The Procter & Gamble Company | Apparatuses and processes for analyzing a system having false start events |
| CN102313577A (en) * | 2011-06-24 | 2012-01-11 | 上海大学 | Equipment health state evaluation and recession prediction method based on multi-channel sensing signals |
| CN103678858A (en) * | 2012-09-26 | 2014-03-26 | 中国人民解放军第二炮兵工程大学 | Method for predicting remaining life of equipment under competing failure conditions |
| CN103868690A (en) * | 2014-02-28 | 2014-06-18 | 中国人民解放军63680部队 | Rolling bearing state automatic early warning method based on extraction and selection of multiple characteristics |
| WO2014161588A1 (en) * | 2013-04-05 | 2014-10-09 | Aktiebolaget Skf | Method for processing data obtained from a condition monitoring system |
| CA2820057A1 (en) * | 2013-07-09 | 2015-01-09 | Tomasz Tomczak | Railing post shoe anchor apparatus and method |
| CN106934125A (en) * | 2017-02-28 | 2017-07-07 | 西安交通大学 | A kind of exponential model plant equipment method for predicting residual useful life of trapezoidal noise profile |
| CN108776017A (en) * | 2018-04-12 | 2018-11-09 | 无锡信捷电气股份有限公司 | A kind of rolling bearing method for predicting residual useful life improving CHSMM |
| CN109299559A (en) * | 2018-10-08 | 2019-02-01 | 重庆大学 | A kind of Surface hardened layer gear wear and fatigue failure competition mechanism analysis method |
| CN109522650A (en) * | 2018-11-16 | 2019-03-26 | 吉林大学 | It is a kind of without burst fail message under electro spindle lifetime estimation method |
| CN110120682A (en) * | 2019-05-08 | 2019-08-13 | 新疆大学 | A kind of minimum tower elevator supply Optimization Scheduling for losing abandonment amount |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4504065B2 (en) * | 2004-03-31 | 2010-07-14 | 中国電力株式会社 | Rolling bearing remaining life diagnosis method |
| EP2475563B1 (en) * | 2009-09-09 | 2014-04-16 | KNORR-BREMSE Systeme für Schienenfahrzeuge GmbH | Method and device for estimating the temperature of an axle bearing of a wheelset of a rail vehicle |
-
2020
- 2020-04-14 CN CN202010290308.XA patent/CN111597682B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5679883A (en) * | 1992-10-19 | 1997-10-21 | Wedeven; Lavern D. | Method and apparatus for comprehensive evaluation of tribological materials |
| WO2001084315A2 (en) * | 2000-05-04 | 2001-11-08 | The Procter & Gamble Company | Apparatus and process for simulating and analyzing a system with failure modes |
| WO2001084254A2 (en) * | 2000-05-04 | 2001-11-08 | The Procter & Gamble Company | Apparatuses and processes for analyzing a system having false start events |
| CN102313577A (en) * | 2011-06-24 | 2012-01-11 | 上海大学 | Equipment health state evaluation and recession prediction method based on multi-channel sensing signals |
| CN103678858A (en) * | 2012-09-26 | 2014-03-26 | 中国人民解放军第二炮兵工程大学 | Method for predicting remaining life of equipment under competing failure conditions |
| WO2014161588A1 (en) * | 2013-04-05 | 2014-10-09 | Aktiebolaget Skf | Method for processing data obtained from a condition monitoring system |
| CA2820057A1 (en) * | 2013-07-09 | 2015-01-09 | Tomasz Tomczak | Railing post shoe anchor apparatus and method |
| CN103868690A (en) * | 2014-02-28 | 2014-06-18 | 中国人民解放军63680部队 | Rolling bearing state automatic early warning method based on extraction and selection of multiple characteristics |
| CN106934125A (en) * | 2017-02-28 | 2017-07-07 | 西安交通大学 | A kind of exponential model plant equipment method for predicting residual useful life of trapezoidal noise profile |
| CN108776017A (en) * | 2018-04-12 | 2018-11-09 | 无锡信捷电气股份有限公司 | A kind of rolling bearing method for predicting residual useful life improving CHSMM |
| CN109299559A (en) * | 2018-10-08 | 2019-02-01 | 重庆大学 | A kind of Surface hardened layer gear wear and fatigue failure competition mechanism analysis method |
| CN109522650A (en) * | 2018-11-16 | 2019-03-26 | 吉林大学 | It is a kind of without burst fail message under electro spindle lifetime estimation method |
| CN110120682A (en) * | 2019-05-08 | 2019-08-13 | 新疆大学 | A kind of minimum tower elevator supply Optimization Scheduling for losing abandonment amount |
Non-Patent Citations (7)
| Title |
|---|
| Li, Peng等.A Competing Risk Model of Reliability Analysis for NAND-Based SSDs in Space Application.IEEE ACCESS.第7卷23430-23441. * |
| Ragab, Ahmed等.Prognostics of multiple failure modes in rotating machinery using a pattern-based classifier and cumulative incidence functions.JOURNAL OF INTELLIGENT MANUFACTURING.2019,第30卷(第1期),255-274. * |
| 李昊.基于随机过程和竞争失效理论的可靠性建模方法.中国优秀硕士学位论文全文数据库工程科技Ⅱ辑.2018,C028-4. * |
| 牛乾.机械旋转部件的性能退化及其寿命预测方法研究.中国博士学位论文全文数据库工程科技Ⅱ辑.2018,C029-20. * |
| 秦荦晟等.小样本下基于竞争失效的轴承可靠性评估.振动与冲击.2017,第36卷(第23期),248-254. * |
| 陆凯等.回转支撑轴承防护罩壳的设计.港口装卸.2020,(第2期),28-31. * |
| 龙哲等.竞争失效的刀具可靠性评估模型.吉林大学学报(工学版).2019,第49卷(第1期),141-148. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111597682A (en) | 2020-08-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111597682B (en) | Method for predicting remaining life of bearing of gearbox of wind turbine | |
| Hsu et al. | Wind turbine fault diagnosis and predictive maintenance through statistical process control and machine learning | |
| Dao et al. | Condition monitoring and fault detection in wind turbines based on cointegration analysis of SCADA data | |
| Cui et al. | An anomaly detection approach based on machine learning and scada data for condition monitoring of wind turbines | |
| CN102022264B (en) | System and method for wind turbine health management | |
| Merkt | On the use of predictive models for improving the quality of industrial maintenance: An analytical literature review of maintenance strategies | |
| CN105205569B (en) | State of fan gear box online evaluation method for establishing model and online evaluation method | |
| Butler et al. | Exploiting SCADA system data for wind turbine performance monitoring | |
| Khodabakhshian | Maintenance management of tractors and agricultural machinery: Preventive maintenance systems | |
| Zhang et al. | A novel fault diagnosis method for wind turbine based on adaptive multivariate time-series convolutional network using SCADA data | |
| Abid et al. | Fault prognostics for the predictive maintenance of wind turbines: State of the art | |
| CN112696481A (en) | Intelligent diagnosis method and device for shaft temperature abnormity of wind turbine generator gearbox | |
| CN112417612A (en) | Method for tracking degradation state and evaluating failure aggregation risk of wind power gear box | |
| CN117195121A (en) | Wind turbine generator abnormal state identification method and system based on improved countermeasure automatic encoder | |
| CN116771610A (en) | Method for adjusting fault evaluation value of variable pitch system of wind turbine | |
| Yongjie et al. | Research on early fault diagnostic method of wind turbines | |
| Dao et al. | Operational condition monitoring of wind turbines using cointegration method | |
| US11339763B2 (en) | Method for windmill farm monitoring | |
| Zhang et al. | Probability warning for wind turbine gearbox incipient faults based on SCADA data | |
| Zhang | Comparison of data-driven and model-based methodologies of wind turbine fault detection with SCADA data | |
| Souza et al. | Evaluation of data based normal behavior models for fault detection in wind turbines | |
| Song et al. | Framework of designing an adaptive and multi-regime prognostics and health management for wind turbine reliability and efficiency improvement | |
| Chacón et al. | SCADA data analytics for fault detection and diagnosis of wind turbines | |
| Ahilan et al. | A Self-Improved Optimizer-Based CNN for Wind Turbine Fault Detection | |
| Wang et al. | Maintenance decision based on data fusion of aero engines |
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 |