CN111783242A - RVM-KF-based rolling bearing residual life prediction method and device - Google Patents
RVM-KF-based rolling bearing residual life prediction method and device Download PDFInfo
- Publication number
- CN111783242A CN111783242A CN202010554878.5A CN202010554878A CN111783242A CN 111783242 A CN111783242 A CN 111783242A CN 202010554878 A CN202010554878 A CN 202010554878A CN 111783242 A CN111783242 A CN 111783242A
- Authority
- CN
- China
- Prior art keywords
- rvm
- data
- prediction
- rolling bearing
- model
- 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
- 238000005096 rolling process Methods 0.000 title claims abstract description 113
- 238000000034 method Methods 0.000 title claims abstract description 64
- 238000006731 degradation reaction Methods 0.000 claims abstract description 64
- 230000015556 catabolic process Effects 0.000 claims abstract description 58
- 238000012544 monitoring process Methods 0.000 claims abstract description 21
- 238000012549 training Methods 0.000 claims abstract description 20
- 230000015654 memory Effects 0.000 claims description 17
- 238000004590 computer program Methods 0.000 claims description 10
- 238000005537 brownian motion Methods 0.000 claims description 4
- 230000008080 stochastic effect Effects 0.000 claims 2
- 230000008569 process Effects 0.000 description 13
- 239000011159 matrix material Substances 0.000 description 7
- 230000000694 effects Effects 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 4
- 230000007774 longterm Effects 0.000 description 4
- 238000012706 support-vector machine Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 238000001914 filtration Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000005653 Brownian motion process Effects 0.000 description 1
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 238000007476 Maximum Likelihood Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 229910021389 graphene Inorganic materials 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000007704 transition 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]
- G06F30/10—Geometric CAD
- G06F30/17—Mechanical parametric or variational design
-
- 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
- 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
- G06F2119/04—Ageing analysis or optimisation against ageing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Geometry (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention relates to a rolling bearing residual life prediction method and device based on RVM-KF. The method comprises the following steps: acquiring real-time monitoring data of a rolling bearing to be predicted and historical data of the whole life cycle of the rolling bearing similar to the rolling bearing to be predicted; training an RVM model by using historical data; determining a prediction starting moment according to the real-time monitoring data and the SKF, and determining a state equation of the Kalman filter according to the real-time monitoring data; inputting the prediction starting time into the trained RVM to obtain estimated vibration data, using the estimated vibration data as an observation value of a Kalman filter to perform single-step prediction, and adding the single-step prediction data into historical data to update the RVM; and performing iterative prediction by using the updated RVM until the vibration data in the single-step prediction data exceeds the failure threshold value to complete the prediction of the residual life. The method combines the RVM model, the Kalman filter and the SKF to realize accurate prediction of the residual life of the rolling bearing in the accelerated degradation stage.
Description
Technical Field
The invention relates to a rolling bearing residual life prediction method and device based on RVM-KF, and belongs to the technical field of fault prediction and health management.
Background
As a basic component widely applied to mechanical equipment, a rolling bearing is one of the key components for determining the safe and reliable operation of the mechanical equipment. According to relevant investigations it was found that approximately 30% of mechanical faults in rotating mechanical equipment are the result of rolling bearing failure. The mechanical failure not only causes economic loss, but also can cause casualties, thereby accurately predicting the residual service life (RUL) of the rolling bearing, and effectively avoiding serious problems of main unit failure, shutdown maintenance, productivity loss, casualties and the like caused by the rolling bearing failure.
At present, the residual life prediction of the rolling bearing is mainly divided into a model-based residual life prediction method and a data-based residual life prediction method.
The model-based residual life prediction method describes the degradation process of the bearing by analyzing the failure mechanism or summarizing experience of the bearing and utilizing a Paris crack propagation model, a Forman crack propagation law, an exponential model, a random model and the like, and then predicts the residual life of the bearing, but the degradation mechanism of the rolling bearing is difficult to obtain accurately and the selection of the model has great influence on the prediction result.
The data-based method does not need to establish a physical model or a statistical model, but obtains the relationship between the degradation state and the residual life of the bearing from the existing monitoring data through models such as a neural network, an SVM (support vector machine), an RVM (relevance vector machine) and the like, thereby predicting the residual life of the rolling bearing.
For this reason, life prediction methods have been proposed that take into account individual differences and overall life characteristics, such as: the journal is 'war industry and engineering journal', the journal is No. 39, No. 5, No. 2018, 5 months, and is named as 'research on a mechanical part residual life prediction model based on a support vector machine and Kalman filtering'. The calculation model can fully utilize the full-life test data of the existing part and the like and the real-time state degradation data of the predicted part to realize the residual life prediction, so that the purpose that the obtained residual life prediction model can utilize the full-life data and considers the individual difference (Kalman filtering model) of a research object is achieved.
However, in order to obtain an accurate nonlinear kalman filter state equation, the method has high requirements on the quality of a training sample and the accuracy of an SVM model, i.e., has high dependence on the model and data, and has low accuracy in the aspect of long-term prediction, so that the method cannot effectively meet the requirements of engineering.
Disclosure of Invention
The application aims to provide a rolling bearing residual life prediction method based on RVM-KF, which is used for solving the problem of inaccurate residual life prediction in the prior art; meanwhile, the rolling bearing residual life prediction device based on the RVM-KF is provided for solving the problem of inaccurate residual life prediction in the prior art.
In order to achieve the purpose, the invention provides a technical scheme of a rolling bearing residual life prediction method based on RVM-KF, which comprises the following steps:
1) acquiring real-time monitoring data of a rolling bearing to be predicted and historical data of the whole life cycle of the rolling bearing similar to the rolling bearing to be predicted; the real-time monitoring data and the historical data refer to vibration data and corresponding time;
2) training an RVM model by using historical data; the input of the RVM model is time, and the output of the RVM model is vibration data;
3) determining a prediction starting time according to the real-time monitoring data and the SKF; determining a state equation of the Kalman filter according to the real-time monitoring data;
4) inputting the prediction starting moment into the trained RVM to obtain estimated vibration data, performing single-step prediction by using the estimated vibration data as an observation value of a Kalman filter to obtain single-step prediction data, finishing prediction of the residual life if the vibration data in the single-step prediction data exceeds a failure threshold, adding the single-step prediction data into the historical data in the step 2) if the vibration data in the single-step prediction data does not exceed the failure threshold, retraining the RVM, and updating the RVM; and the step is executed iteratively until the vibration data in the single step prediction data exceeds a failure threshold value.
In addition, the invention also provides a technical scheme of the rolling bearing residual life prediction device based on the RVM-KF, which comprises a processor, a memory and a computer program which is stored in the memory and can run on the processor, wherein the processor realizes the technical scheme of the rolling bearing residual life prediction method based on the RVM-KF when executing the computer program.
The technical scheme of the rolling bearing residual life prediction method and device based on RVM-KF has the beneficial effects that: according to the method, an RVM model is established through historical data of similar rolling bearings, the time when the rolling bearings to be predicted enter accelerated degradation is determined according to real-time monitoring data and SKF of the rolling bearings to be predicted, the prediction starting time is further determined, meanwhile, a state equation of a Kalman filter is determined according to the real-time monitoring data, vibration data are estimated according to the prediction starting time and the established RVM model, the estimated values are used as observed values of the Kalman filter to perform single-step prediction, when the vibration data predicted by a single step do not exceed a failure threshold, the predicted data are added into the RVM model, the RVM model is retrained, iterative prediction is performed until the vibration data predicted by the single step exceed the failure threshold, the iteration is finished, and the residual life is obtained. According to the invention, the RVM model is continuously updated, the RVM model, the Kalman filter and the SKF are combined to realize accurate prediction of the residual life of the rolling bearing in the accelerated degradation stage, the uncertainty of long-term prediction of the residual life of the rolling bearing is reduced, the long-term prediction precision of the residual life of the rolling bearing is improved, and the Kalman filter and the state equation are combined to effectively reduce the dependence of the prediction process on data and the model.
Furthermore, in the rolling bearing residual life prediction method and device based on the RVM-KF, in order to better fit the accelerated degradation process, the state equation of the Kalman filter is a random effect index model.
Further, in the rolling bearing residual life prediction method and device based on the RVM-KF, the random effect index model is:
wherein Y (t) is vibration data at t moment, α is constant, theta and β are parameters, and sigma is2Is the variance; (t) is an error term with a mean of 0 and a variance of σ2Brownian motion of.
Further, in the rolling bearing residual life prediction method and device based on the RVM-KF, the vibration data comprises at least one characteristic value extracted from the vibration signal.
Furthermore, in the rolling bearing residual life prediction method and device based on the RVM-KF, in order to improve the prediction accuracy and reduce the complexity of model training, the historical data used in the training of the RVM model is the historical data of the accelerated degradation stage.
Drawings
FIG. 1 is a flow chart of the method for predicting the residual life of a rolling bearing based on RVM-KF in accordance with the present invention;
FIG. 2 is a degradation curve of the rolling bearing of the present invention;
FIG. 3 is a graph comparing the predicted degradation trajectory of the rolling bearing 2-3 of the present invention with the predicted degradation trajectory of the rolling bearing 2-3 of the prior art;
FIG. 4 is a graph comparing the predicted degradation trajectory of the rolling bearing 2-5 of the present invention with the predicted degradation trajectory of the rolling bearing 2-5 of the prior art;
FIG. 5 is a graph comparing the predicted residual life of the rolling bearing 2-3 of the present invention with the predicted residual life of the rolling bearing 2-3 of the prior art;
FIG. 6 is a graph comparing the predicted residual life of the rolling bearing 2-5 of the present invention with the predicted residual life of the rolling bearing 2-5 of the prior art;
FIG. 7 is a schematic structural diagram of the rolling bearing residual life prediction device based on RVM-KF in accordance with the present invention.
Detailed Description
Embodiment of the rolling bearing residual life prediction method based on RVM-KF:
the RVM-KF-based rolling bearing residual life prediction method has the main concept that the RVM model, the Kalman filter and the SKF are combined to iteratively predict the residual life of the rolling bearing entering the accelerated degradation stage, so that the accuracy of life prediction is improved.
Specifically, the method for predicting the residual life of the rolling bearing based on the RVM-KF is shown in FIG. 1 and comprises the following steps:
1) historical data of the whole life cycle of the rolling bearing of the same type as the rolling bearing to be predicted is collected, and the historical data comprises vibration signals and corresponding time (the time is the use time).
In order to reduce individual differences, different model parameters are applied to rolling bearings of different models for service life prediction, so that the rolling bearings corresponding to historical use data selected during modeling should be the same as the rolling bearings to be predicted (namely the same model).
The method comprises the steps of carrying out a full-life cycle test on a plurality of rolling bearings of the same type under the same test conditions, periodically acquiring vibration signals (acquiring vibration signals of 1.28s every 1 min) in the working process of the rolling bearings in the test process until the rolling bearings completely fail, and obtaining the full-life vibration signals of the plurality of rolling bearings, wherein each vibration signal has corresponding service time.
2) Extracting a characteristic value which can effectively represent the degradation of the rolling bearing from the vibration signal of the step 1).
In order to predict the remaining life, a characteristic value representing degradation of the rolling bearing is extracted from the vibration signal as vibration data, however, the vibration signal includes a plurality of kinds of vibration data, and in a case that the vibration data representing degradation can be obtained by calculation with other vibration data, other vibration data can also be extracted, which is not limited by the present invention.
3) Inputting the characteristic values obtained in the step 2) and the corresponding time of each characteristic value in the step 1) into a pre-established SKF model, and finding out the accelerated degradation time (namely the early failure point) of the similar rolling bearing in the life cycle.
The degradation process of the rolling bearing comprises a stable degradation stage and an accelerated degradation stage as shown in fig. 2, the SKF model is a dynamic bayesian network and comprises two kalman filters, the two kalman filters comprise a first kalman filter for predicting the probability of stable degradation at a certain moment, and the first kalman filter is suitable for the stable degradation stage; and a second Kalman filter for predicting the probability of accelerated degradation at a certain moment, and is suitable for the accelerated degradation stage. The data input by the SKF can obtain a first probability that each moment belongs to a stable degradation stage through a first Kalman filter, and can obtain a second probability that each moment belongs to an accelerated degradation stage through a second Kalman filter.
The specific probability calculation process of the kalman filter is as follows:
the first Kalman filter and the second Kalman filter are basic Kalman filters, and the prediction and update processes are as follows:
wherein, ytFor the characteristic value monitored at the time t,for the characteristic value estimated at time t, phitFor the state transition matrix at time t,is phitTransposed matrix of (H)tFor the observation matrix at the time t,is HtThe transpose matrix of (a) is,tas measured residual at time t, CtIs the residual covariance at time t, Ct -1Is CtInverse matrix of, KtKalman gain, Q, at time ttIs the variance of the process noise at time t, RtThe variance of the noise is measured for time t,is an estimate of the covariance, I is the identity matrix, xtEstimated characteristic value, P, updated for time ttAn updated estimate of the covariance for time t.
Then, the measurement residuals of the ith Kalman filter follow the probability density function of normal distribution with the mean value of 0Comprises the following steps:
the probability of each kalman filter is obtained as follows:
Wherein the content of the first and second substances,is the probability that the characteristic value belongs to the ith Kalman filter at the moment t, ZijThe probability of transferring the ith Kalman filter to the jth Kalman filter, n is the number of Kalman filters,for the measurement residual of the i-th kalman filter at time t,is the residual covariance of the ith kalman filter at time t,for the ith card at time tUpdated state of the Kalman filter, Pt iCovariance updated for the ith kalman filter at time t. And substituting the calculated state weight and covariance weight into the Kalman filtering process to obtain the most possible degradation state at each moment.
After the first probability and the second probability of each time are obtained, if the second probability of a plurality of continuous times (the plurality of times are generally 5 times, and certainly the number of the plurality of times can be set as required) is greater than the first probability, the foremost time of the plurality of times is used as the accelerated degradation time, and as another embodiment, the accelerated degradation time can also be calculated according to the plurality of times, or the rearmost time of the plurality of times is used as the accelerated degradation time, which is not limited by the present invention.
As another embodiment, other methods may be used to determine the stage of the rolling bearing to be predicted, for example: the stage can also be determined by comparing the characteristic value with the empirical value. The invention does not limit the way how the stage of the rolling bearing to be predicted is judged.
4) And 3) training the RVM model by taking the historical data after the accelerated degradation moment obtained in the step 3) as a training data set.
The RVM model is a generalized linear model under a Bayesian framework, and the training data set isThe RVM model established is as follows:
the likelihood function of the training samples is:
wherein lnIs the nth target data (i.e., output data), here, the eigenvalues;tnis the nth input data, here time (i.e., duration of use); n is the total number of data in the training data set; omeganThe weight corresponding to the nth data, K (t, t)n) A kernel function corresponding to the nth data;nis a mean value of 0 and a variance of σ2Phi is an N × N +1 matrix based on kernel function, phi is [ phi (t)1),φ(t2),…,φ(tn),…,φ(tN)]T,φ(tn) Is the kernel vector corresponding to the nth data, phi (t)n)=[1,K(tn,t1),K(tn,t2),…K(tn,tN)]K(tn,t1),K(tn,t2),…,K(tn,tN) Are both K (t, t)n) Vector element in (1), t is input set; l is an output set; ω is a weight set, ω ═ ω (ω ═ ω)1,ω2,…,ωn,…ωN)T。
Using maximum likelihood method directly to pair omega and sigma2The result is that the elements in ω are usually mostly not 0, resulting in an overfitting. To avoid this in RVM, we add a prerequisite to ω: their probability distribution is a normal distribution falling around 0:
wherein α is a vector of N +1, αiIs the ith hyperparameter; n () is a normal distribution.
Thus, for a newly given input t*Its predicted value l*:
p(l*|l)=∫p(l*|ω,α,σ2)p(ω,α,σ2|l)dωdαdσ2。
In this step, historical data after the accelerated degradation moment is optimized for the training of the RVM model, so that not only excessive data amount can be avoided, but also the RVM model can be estimated for the accelerated degradation stage, and the estimation is more accurate.
5) Acquiring real-time monitoring data of a rolling bearing to be predicted, wherein the real-time monitoring data comprises a vibration signal and corresponding time (here, using time); and extracting a real-time characteristic value capable of effectively representing the degradation of the rolling bearing from the vibration signal, judging the stage of the rolling bearing to be predicted according to the real-time characteristic value and the SKF, and determining the prediction starting time according to the accelerated degradation time if the rolling bearing to be predicted enters the accelerated degradation time.
The predicted start time is typically slightly later than the accelerated degradation time, for example: the accelerated degradation time is the time corresponding to 200min of use, and then the predicted starting time is the time corresponding to 220min of use, and of course, the predicted starting time can be set as required, and the invention is not limited.
The process of judging the stage of the rolling bearing to be predicted is the same as that in the step 3), except that the data of the SKF is input as a real-time characteristic value, and therefore the specific judgment process is not repeated here.
6) And determining parameters of a state equation in a Kalman Filter (KF) according to the real-time monitoring data, and determining the state equation, wherein the state equation is a random effect index model.
The real-time monitoring data for determining the state equation parameters may be monitoring data near the accelerated degradation time, and may be monitoring data near the predicted starting time, which is not limited in the present invention. The state equation may be updated according to the update of the monitoring data until the prediction start time is not determined, but the state equation is unchanged after the prediction start time is determined.
In this embodiment, the degradation curve of the rolling bearing is as shown in fig. 2, and includes a stable degradation state and an accelerated degradation state, and as can be seen from the graph, the accelerated degradation curve is similar to an exponential function curve, so that the state equation of the kalman filter is represented by a random effect exponential model, and the random coefficient exponential model is:
where Y (t) is vibration data at time t (here, a characteristic value of the rolling bearing), α is a constant, theta and β are parameters (i.e., random parameters), and (t) is an error term having a mean value of 0 and a variance of σ2Brownian motion of.
For convenience of calculation, the random coefficient index model is logarithmized and transformed into:
in the above formula: theta' is equal to ln theta and is,theta 'and β' are parameters after deformation, and (t) represents error, and is a mean value of 0 and a variance of sigma2The central brownian motion of.
As another embodiment, the state equation of the third kalman filter may be any existing model, and may conform to the process of accelerated degradation, which is not limited in the present invention.
7) Inputting the prediction starting time determined in the step 5) into the RVM model trained in the step 4) to obtain an estimated characteristic value of the rolling bearing to be predicted, taking the estimated characteristic value of the rolling bearing to be predicted as a measurement value of a Kalman filter, performing single-step prediction to obtain single-step prediction data, adding the single-step prediction data into the training data set in the step 4) to retrain the RVM model, and updating the RVM model.
The single step prediction data includes the time of the single step prediction: predicting the starting time + step length and the characteristic value of single-step prediction;
8) and estimating the characteristic value of the next moment by using the updated RVM model, and performing iterative prediction until the characteristic value (namely the predicted characteristic value) in the single-step prediction data exceeds a failure threshold value, thereby completing the prediction of the residual life.
And 8) judging whether the vibration data in the single-step prediction data exceed a failure threshold value, if not, continuing iterative prediction, and if so, determining the residual life of the rolling bearing to be predicted according to the iteration times.
The RVM model is estimated every time it is updated in step 8), for example, the feature value of the next time of use: determining the initial prediction time as 220min, obtaining an estimated characteristic value used for 220min according to the RVM trained in the step 4), and performing single-step prediction on the estimated characteristic value used for 220min through KF to obtain single-step prediction data; the single step prediction data comprises the use of 220min and a corresponding prediction characteristic value; if the prediction characteristic value corresponding to the use time of 220min does not exceed the failure threshold value, adding the single-step prediction data into the training data set in the step 4), retraining the RVM model, and updating the RVM model;
then, estimating a characteristic value using 220min +1min according to the updated RVM model to obtain an estimated characteristic value using 221min, and performing single-step prediction on the estimated characteristic value using 221min through KF to obtain single-step prediction data; the single-step prediction data of the iteration comprises 221min and a corresponding prediction characteristic value;
and by analogy, performing iterative prediction, and if the predicted characteristic value in the single-step prediction data exceeds the failure threshold value after the characteristic value of 230min is used for estimation, determining that the residual life is 230-220 min-10 min.
In order to obtain the residual life corresponding to different prediction starting moments, after the life prediction is finished at the first prediction starting moment, the prediction starting moment is updated to perform the life prediction again, at the moment, a state equation is also required to be updated, namely, the steps 6 to 8 are repeated), and the prediction of the residual life at the new prediction starting moment is completed. And further obtaining the residual service life corresponding to different predicted starting moments.
The validity and correctness of the present invention will be verified by combining the examples below, wherein the data set XJTU-SY is provided by the institute of design science and infrastructure of the university of West-Ann transportation (XJTU) and the Yangtze science and technology corporation (SY) of Changxing Shang Yang in Zhejiang.
The invention divides the degradation process of the rolling bearings into stable degradation and accelerated degradation, extracts the wavelet packet coefficient root mean square (characteristic value) of 5 rolling bearings under the working condition 2 to identify the degradation state, the serial numbers of the 5 rolling bearings are respectively 2-1, 2-2, 2-3, 2-4 and 2-5, and the time when each rolling bearing enters the accelerated degradation stage is shown as table one. The failure threshold value of a rolling bearing is set to be 13, the rolling bearings with the serial numbers of 2-3 and 2-5 are used as rolling bearings to be predicted, and when the serial numbers of 2-3 are used as the rolling bearings to be predicted, the data of the rolling bearings with the serial numbers of 2-1, 2-2, 2-4 and 2-5 are used as a training set; when the serial number 2-5 is taken as the rolling bearing to be predicted, the data of the rolling bearings with the serial numbers 2-1, 2-2, 2-3 and 2-4 are taken as a training set; and (3) utilizing the wavelet packet coefficient root mean square of the other 4 rolling bearings in the accelerated degradation stage as an RVM training set, estimating the degradation characteristic value of the rolling bearing to be predicted according to the trained RVM model, performing single-step prediction processing by taking the estimated value of the RVM as the measured value of KF, adding the KF filter value into the RVM training set to retrain the RVM model for next iterative prediction until the failure threshold of the rolling bearing is reached, and completing prediction of the residual life of the bearing. The degradation locus and the residual life of the bearing estimated by the RVM, the KF and the iterative method are compared, and the comparison results are shown in figures 3, 4, 5 and 6. And as shown in table two, the accuracy of the degraded track prediction by 3 methods is given by using the absolute error MAE and the root mean square error RMSE.
in the formula, xiIs the true degradation characteristic quantity, y, of the bearingiFor the predicted degradation feature data, n is the total number of data.
TABLE 5 accelerated degradation moments of rolling bearings
Bearing number | 2-1 | 2-2 | 2-3 | 2-4 | 2-5 |
Entering accelerated degradation stage moment (min) | 456 | 55 | 305 | 31 | 124 |
The rolling bearings with the serial numbers 2-3 and 2-5 in the table adopt different absolute errors and root mean square errors of prediction methods
As can be seen from the comparison of FIGS. 3 and 4, the predicted result of the method (RVM-KF) of the present invention is closer to the true value. The method is based on the SKF model, utilizes the nonlinear degradation of the piecewise linear approximation rolling bearing to track the dynamic change of the rolling bearing degradation process, does not need to set a threshold value and a large amount of data training model, and adaptively obtains the most possible degradation model at each moment; the RVM is a generalized linear model under a Bayesian framework, is more suitable for processing small sample problems compared with a neural network RVM, and can provide complete prediction distribution; meanwhile, the KF and the random effect index model are combined, so that the degradation process of the bearing can be better fitted, and the service life is more accurately predicted.
The method can accurately predict the residual service life of the bearing at the early stage of the accelerated degradation stage in real time without an accurate model and a large amount of high-quality data, and simultaneously reduces the uncertainty of long-term prediction.
RVM-KF-based rolling bearing residual life prediction device embodiment:
the rolling bearing residual life prediction device based on the RVM-KF comprises a processor, a memory and a computer program which is stored in the memory and can run on the processor, wherein the processor realizes the rolling bearing residual life prediction method based on the RVM-KF when executing the computer program, as shown in FIG. 7.
The specific implementation process and effect of the rolling bearing residual life prediction method based on the RVM-KF are described in the embodiment of the rolling bearing residual life prediction method based on the RVM-KF, and are not described herein again.
That is, the method in the above RVM-KF based rolling bearing remaining life prediction method embodiment should understand that the flow of the RVM-KF based rolling bearing remaining life prediction method may be realized by computer program instructions. These computer program instructions may be provided to a processor (e.g., a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus), such that the instructions, which execute via the processor, create means for implementing the functions specified in the method flow.
The processor referred to in this embodiment refers to a processing device such as a microprocessor MCU or a programmable logic device FPGA;
the memory of the present embodiment is used for storing computer program instructions for implementing the rolling bearing remaining life prediction method based on RVM-KF, and includes a physical device for storing information, and the information is usually digitized and then stored in a medium using an electric, magnetic, or optical method. For example: various memories for storing information by using an electric energy mode, such as RAM, ROM and the like; various memories for storing information by magnetic energy, such as hard disk, floppy disk, magnetic tape, magnetic core memory, bubble memory, and U disk; various types of memory, CD or DVD, that store information optically. Of course, there are other ways of memory, such as quantum memory, graphene memory, and so forth.
The rolling bearing residual life prediction device based on the RVM-KF, which is formed by the memory and the processor, wherein the memory is used for storing computer program instructions formed by realizing the rolling bearing residual life prediction method based on the RVM-KF, and the processor executes corresponding program instructions in the computer, and the computer can be realized by using a windows operating system, a linux system or other systems, for example, an android system programming language, an iOS system programming language, a processing logic realization based on a quantum computer, and the like.
As another embodiment, the rolling bearing remaining life prediction device based on the RVM-KF may further include other processing hardware, such as a database, a multi-level cache, a GPU, and the like.
Claims (6)
1. A rolling bearing residual life prediction method based on RVM-KF is characterized by comprising the following steps:
1) acquiring real-time monitoring data of a rolling bearing to be predicted and historical data of the whole life cycle of the rolling bearing similar to the rolling bearing to be predicted; the real-time monitoring data and the historical data refer to vibration data and corresponding time;
2) training an RVM model by using historical data; the input of the RVM model is time, and the output of the RVM model is vibration data;
3) determining a prediction starting time according to the real-time monitoring data and the SKF; determining a state equation of the Kalman filter according to the real-time monitoring data;
4) inputting the prediction starting moment into the trained RVM to obtain estimated vibration data, taking the estimated vibration data as an observation value of a Kalman filter to perform single-step prediction to obtain single-step prediction data, finishing prediction of the residual life if the vibration data in the single-step prediction data exceeds a failure threshold, adding the single-step prediction data into the historical data in the step 2) if the vibration data in the single-step prediction data does not exceed the failure threshold, retraining the RVM, and updating the RVM; and the step is executed iteratively until the vibration data in the single step prediction data exceeds a failure threshold value.
2. The RVM-KF-based rolling bearing residual life prediction method of claim 1, characterized in that the Kalman filter's equation of state is a stochastic effect exponential model.
3. The RVM-KF based rolling bearing residual life prediction method of claim 2 wherein said stochastic effect index model is:
wherein Y (t) is vibration data at t moment, α is constant, theta and β are parameters, and sigma is2Is the variance; (t) is an error term with a mean of 0 and a variance of σ2Brownian motion of.
4. The RVM-KF based rolling bearing remaining life prediction method of claim 1 wherein said vibration data includes at least one characteristic value extracted from the vibration signal.
5. The RVM-KF-based rolling bearing residual life prediction method of claim 1 wherein the historical data utilized in training the RVM model is the historical data of the accelerated degradation phase.
6. An RVM-KF based rolling bearing remaining life predicting device, characterized by comprising a processor, a memory and a computer program stored in the memory and executable on the processor, the processor implementing the RVM-KF based rolling bearing remaining life predicting method according to any one of claims 1-5 when executing the computer program.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010554878.5A CN111783242B (en) | 2020-06-17 | 2020-06-17 | RVM-KF-based rolling bearing residual life prediction method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010554878.5A CN111783242B (en) | 2020-06-17 | 2020-06-17 | RVM-KF-based rolling bearing residual life prediction method and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111783242A true CN111783242A (en) | 2020-10-16 |
CN111783242B CN111783242B (en) | 2024-05-28 |
Family
ID=72757291
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010554878.5A Active CN111783242B (en) | 2020-06-17 | 2020-06-17 | RVM-KF-based rolling bearing residual life prediction method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111783242B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112949204A (en) * | 2021-03-22 | 2021-06-11 | 西安交通大学 | Rolling bearing residual life prediction method with data and model adaptive matching |
CN113609685A (en) * | 2021-08-09 | 2021-11-05 | 电子科技大学 | Bearing residual life prediction method based on optimized RVM and mixed degradation model |
CN114662226A (en) * | 2022-02-24 | 2022-06-24 | 北京工业大学 | Time-varying Kalman filtering-based method for predicting residual service life of rolling bearing |
Citations (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1705542A1 (en) * | 2005-03-24 | 2006-09-27 | Abb Research Ltd. | Estimating health parameters or symptoms of a degrading system |
CN104318043A (en) * | 2014-02-20 | 2015-01-28 | 河南科技大学 | Rolling bearing vibration performance reliability variation process detection method and rolling bearing vibration performance reliability variation process detection device |
CN104615866A (en) * | 2015-01-21 | 2015-05-13 | 北京航空航天大学 | Service life prediction method based on physical statistic model |
CN105300692A (en) * | 2015-08-07 | 2016-02-03 | 浙江工业大学 | Bearing fault diagnosis and prediction method based on extended Kalman filtering algorithm |
CN105956236A (en) * | 2016-04-22 | 2016-09-21 | 西安交通大学 | Dual-updating four-factor random degeneration model gear life prediction method |
KR20170067292A (en) * | 2015-12-08 | 2017-06-16 | 한양대학교 산학협력단 | Device and method for estimating remaining life of mechanical system |
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 |
CN107145720A (en) * | 2017-04-19 | 2017-09-08 | 浙江大学 | It is continuous to degenerate and the unknown equipment method for predicting residual useful life impacted under collective effect |
CN108846241A (en) * | 2018-07-04 | 2018-11-20 | 西安交通大学 | The rolling bearing life prediction technique to be linked based on Fu Leixie apart from dynamic digital-to-analogue |
CN109543317A (en) * | 2018-04-28 | 2019-03-29 | 北京航空航天大学 | A kind of method and device of PEMFC remaining life prediction |
CN109726524A (en) * | 2019-03-01 | 2019-05-07 | 哈尔滨理工大学 | A kind of rolling bearing remaining life prediction technique based on CNN and LSTM |
CN110095731A (en) * | 2019-05-06 | 2019-08-06 | 哈尔滨工业大学 | Remaining life DIRECT FORECASTING METHOD applied to long-life space lithium ion battery |
US20200116553A1 (en) * | 2018-10-12 | 2020-04-16 | Industrial Technology Research Institute | Health monitor method for an equipment and system thereof |
US20200116585A1 (en) * | 2018-10-10 | 2020-04-16 | Palo Alto Research Center Incorporated | Adaptive remaining useful life estimation method using constraint convex regression from degradation measurement |
KR20200049295A (en) * | 2018-10-31 | 2020-05-08 | 한국전력공사 | A method to predict health index transition and residual life for turbomachinery |
-
2020
- 2020-06-17 CN CN202010554878.5A patent/CN111783242B/en active Active
Patent Citations (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1705542A1 (en) * | 2005-03-24 | 2006-09-27 | Abb Research Ltd. | Estimating health parameters or symptoms of a degrading system |
CN104318043A (en) * | 2014-02-20 | 2015-01-28 | 河南科技大学 | Rolling bearing vibration performance reliability variation process detection method and rolling bearing vibration performance reliability variation process detection device |
CN104615866A (en) * | 2015-01-21 | 2015-05-13 | 北京航空航天大学 | Service life prediction method based on physical statistic model |
CN105300692A (en) * | 2015-08-07 | 2016-02-03 | 浙江工业大学 | Bearing fault diagnosis and prediction method based on extended Kalman filtering algorithm |
KR20170067292A (en) * | 2015-12-08 | 2017-06-16 | 한양대학교 산학협력단 | Device and method for estimating remaining life of mechanical system |
CN105956236A (en) * | 2016-04-22 | 2016-09-21 | 西安交通大学 | Dual-updating four-factor random degeneration model gear life prediction 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 |
CN107145720A (en) * | 2017-04-19 | 2017-09-08 | 浙江大学 | It is continuous to degenerate and the unknown equipment method for predicting residual useful life impacted under collective effect |
CN109543317A (en) * | 2018-04-28 | 2019-03-29 | 北京航空航天大学 | A kind of method and device of PEMFC remaining life prediction |
CN108846241A (en) * | 2018-07-04 | 2018-11-20 | 西安交通大学 | The rolling bearing life prediction technique to be linked based on Fu Leixie apart from dynamic digital-to-analogue |
US20200116585A1 (en) * | 2018-10-10 | 2020-04-16 | Palo Alto Research Center Incorporated | Adaptive remaining useful life estimation method using constraint convex regression from degradation measurement |
US20200116553A1 (en) * | 2018-10-12 | 2020-04-16 | Industrial Technology Research Institute | Health monitor method for an equipment and system thereof |
KR20200049295A (en) * | 2018-10-31 | 2020-05-08 | 한국전력공사 | A method to predict health index transition and residual life for turbomachinery |
CN109726524A (en) * | 2019-03-01 | 2019-05-07 | 哈尔滨理工大学 | A kind of rolling bearing remaining life prediction technique based on CNN and LSTM |
CN110095731A (en) * | 2019-05-06 | 2019-08-06 | 哈尔滨工业大学 | Remaining life DIRECT FORECASTING METHOD applied to long-life space lithium ion battery |
Non-Patent Citations (6)
Title |
---|
KE LI ET AL.: "New Particle Filter Based on GA for Equipment Remaining Useful Life Prediction", SENSORS, vol. 17, no. 4, 28 March 2017 (2017-03-28) * |
YUCHEN SONG ET AL.: "Satellite lithium-ion battery remaining useful life estimation with an iterative updated RVM fused with the KF algorithm", CHINESE JOURNAL OF AERONAUTICS, vol. 31, no. 1 * |
于震梁;孙志礼;曹汝男;王鹏;: "基于支持向量机和卡尔曼滤波的机械零件剩余寿命预测模型研究", 兵工学报, no. 05, 31 May 2018 (2018-05-31) * |
夏新涛等: "滚动轴承性能保持可靠性预测", 轴承, no. 6 * |
彭开香;皮彦婷;焦瑞华;唐鹏;: "航空发动机的健康指标构建与剩余寿命预测", 控制理论与应用, no. 04 * |
阙子俊;金晓航;孙毅;: "基于UKF的轴承剩余寿命预测方法研究", 仪器仪表学报, no. 09 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112949204A (en) * | 2021-03-22 | 2021-06-11 | 西安交通大学 | Rolling bearing residual life prediction method with data and model adaptive matching |
CN112949204B (en) * | 2021-03-22 | 2022-12-09 | 西安交通大学 | Rolling bearing residual life prediction method with data and model adaptive matching |
CN113609685A (en) * | 2021-08-09 | 2021-11-05 | 电子科技大学 | Bearing residual life prediction method based on optimized RVM and mixed degradation model |
CN113609685B (en) * | 2021-08-09 | 2023-04-07 | 电子科技大学 | Bearing residual life prediction method based on optimized RVM and mixed degradation model |
CN114662226A (en) * | 2022-02-24 | 2022-06-24 | 北京工业大学 | Time-varying Kalman filtering-based method for predicting residual service life of rolling bearing |
CN114662226B (en) * | 2022-02-24 | 2024-03-29 | 北京工业大学 | Rolling bearing residual service life prediction method based on time-varying Kalman filtering |
Also Published As
Publication number | Publication date |
---|---|
CN111783242B (en) | 2024-05-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112629863B (en) | Bearing fault diagnosis method for dynamic joint distribution alignment network under variable working conditions | |
CN108984893B (en) | Gradient lifting method-based trend prediction method | |
CN111783242A (en) | RVM-KF-based rolling bearing residual life prediction method and device | |
CN111414703B (en) | Method and device for predicting residual life of rolling bearing | |
CN113722985B (en) | Method and system for evaluating health state and predicting residual life of aero-engine | |
CN112433896B (en) | Method, device, equipment and storage medium for predicting server disk faults | |
CN111625516A (en) | Method and device for detecting data state, computer equipment and storage medium | |
CN108804334B (en) | Discrete software reliability increase testing and evaluating method based on self-adaptive sampling | |
CN103310051A (en) | Method for predicting failure rate of vehicle-mounted information terminal in entire life cycle | |
CN108961460B (en) | Fault prediction method and device based on sparse ESGP (Enterprise service gateway) and multi-objective optimization | |
Zhang et al. | A novel data-driven method based on sample reliability assessment and improved CNN for machinery fault diagnosis with non-ideal data | |
CN110610226A (en) | Generator fault prediction method and device | |
CN114662386A (en) | Bearing fault diagnosis method and system | |
CN114154252B (en) | Risk assessment method and device for failure mode of power battery system of new energy automobile | |
Ge et al. | An improved PF remaining useful life prediction method based on quantum genetics and LSTM | |
CN117521512A (en) | Bearing residual service life prediction method based on multi-scale Bayesian convolution transducer model | |
CN111523727A (en) | Method for predicting remaining life of battery by considering recovery effect based on uncertain process | |
CN108932197B (en) | Software failure time prediction method based on parameter Bootstrap resampling | |
CN111400964B (en) | Fault occurrence time prediction method and device | |
CN113537614A (en) | Construction method, system, equipment and medium of power grid engineering cost prediction model | |
CN113822336A (en) | Cloud hard disk fault prediction method, device and system and readable storage medium | |
CN111241629B (en) | Intelligent prediction method for performance change trend of hydraulic pump of airplane based on data driving | |
CN116303786B (en) | Block chain financial big data management system based on multidimensional data fusion algorithm | |
Wang et al. | Similarity-based echo state network for remaining useful life prediction | |
CN112395167A (en) | Operation fault prediction method and device and electronic equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |