CN112100919A - Rolling bearing residual life prediction method based on RE-CF-EKF algorithm - Google Patents
Rolling bearing residual life prediction method based on RE-CF-EKF algorithm Download PDFInfo
- Publication number
- CN112100919A CN112100919A CN202010968660.4A CN202010968660A CN112100919A CN 112100919 A CN112100919 A CN 112100919A CN 202010968660 A CN202010968660 A CN 202010968660A CN 112100919 A CN112100919 A CN 112100919A
- Authority
- CN
- China
- Prior art keywords
- prediction
- residual life
- characteristic curve
- ekf algorithm
- bearing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 48
- 238000000034 method Methods 0.000 title claims abstract description 48
- 238000005096 rolling process Methods 0.000 title claims abstract description 27
- 238000012549 training Methods 0.000 claims abstract description 21
- 238000012360 testing method Methods 0.000 claims abstract description 20
- 238000001914 filtration Methods 0.000 claims abstract description 14
- 238000012216 screening Methods 0.000 claims abstract description 4
- 238000009825 accumulation Methods 0.000 claims description 11
- 230000015556 catabolic process Effects 0.000 claims description 9
- 238000006731 degradation reaction Methods 0.000 claims description 9
- 230000001186 cumulative effect Effects 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 7
- 230000009467 reduction Effects 0.000 claims description 7
- 238000005259 measurement Methods 0.000 claims description 6
- 238000000354 decomposition reaction Methods 0.000 claims description 4
- 101000802640 Homo sapiens Lactosylceramide 4-alpha-galactosyltransferase Proteins 0.000 claims description 3
- 102100035838 Lactosylceramide 4-alpha-galactosyltransferase Human genes 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 230000000694 effects Effects 0.000 description 6
- 238000004088 simulation Methods 0.000 description 5
- 238000011161 development Methods 0.000 description 4
- 230000036541 health Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000000704 physical effect Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 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
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/04—Ageing analysis or optimisation against ageing
Abstract
The invention provides a rolling bearing residual life prediction method based on an RE-CF-EKF algorithm, which is characterized by comprising the following steps: the method comprises the following steps: screening the obtained bearing vibration data set, selecting a plurality of bearings under the same working condition as an experimental data set, and grouping the experimental data set with the bearings under the same working condition into a training set and a test set; step two: linear filtering is carried out after the time domain characteristics of the original signal are taken, a relative error value RE between filtered data and a filtering error is calculated, and a prediction starting point TSP is determined based on an RE curve; step three: denoising the selected time domain characteristics, and obtaining an accumulated characteristic CF with smoother trend by an accumulated function; step four: and combining the second step and the third step to obtain a characteristic curve section suitable for modeling prediction, establishing a state space model, obtaining model initial parameters from the characteristic curve section of the training set, and then combining an EKF algorithm to predict the residual life of the bearing.
Description
Technical Field
The invention belongs to the technical field of rolling bearings, and relates to a rolling bearing residual life prediction method based on an RE-CF-EKF algorithm.
Background
Today, the pace of economic development is rapid, the development level of science and technology has been taken as an important sign of the comprehensive strength of the country, wherein the development of the manufacturing industry has a great influence on the development of the science and technology of the country. As one of the key components of a rotating machine, a bearing whose smooth and reliable operation is essential to maintain production. Unfortunately, bearing failure is one of the most common failure modes in machine failure. If the health condition of the rolling bearing can be evaluated and estimated in the working process of the rolling bearing, the service life of the bearing can be utilized to the maximum extent before safety is ensured, and the maintenance and guarantee cost is reduced. In recent years, a lot of research has been conducted in the field of Prognostic Health Management (PHM) both domestically and abroad, and the first method is a physical model-based method, which relies on a priori knowledge of the intrinsic system failure mechanism to construct a degradation model, thereby describing the physical properties of the failure. The second method is a data-driven method, which usually does not need to know the physical properties of equipment failure and degradation, and only needs to use a large amount of data to find the degradation rule of the equipment. Chang proposes a mixed method of error correction ideas to predict the residual service life, integrates unscented Kalman filtering algorithm (UKF), complete set empirical mode decomposition (CEEMD) and relevance vector machine (RVR) methods to realize RUL prediction, but does not consider the prediction starting point and more trending characteristics.
Disclosure of Invention
The invention aims to overcome the defects and provides a method for accurately predicting the residual service life of a rolling bearing. The method adopts the RE-CF-EKF algorithm, can realize accurate detection of the degradation starting point, reduces the influence of potential health factors on fault factor prediction, and simultaneously utilizes the accumulation feature (CF) with stronger trend after accumulation function processing, can effectively improve the EKF prediction precision and realize accurate residual life prediction result.
The invention provides a rolling bearing residual life prediction method based on an RE-CF-EKF algorithm, which is characterized by comprising the following steps:
the method comprises the following steps: screening the obtained bearing vibration data set, selecting a plurality of bearings under the same working condition as an experimental data set, and grouping the experimental data set with the bearings under the same working condition into a training set and a test set;
step two: linear filtering is carried out after the time domain characteristics of the original signal are taken, a relative error value RE between filtered data and a filtering error is calculated, and a prediction starting point TSP is determined based on an RE curve;
step three: denoising the selected time domain characteristics, and obtaining an accumulated characteristic CF with smoother trend by an accumulated function;
step four: and combining the second step and the third step to obtain a characteristic curve section suitable for modeling prediction, establishing a state space model, obtaining model initial parameters from the characteristic curve section of the training set, and then combining an EKF algorithm to predict the residual life of the bearing.
Further, the invention provides a rolling bearing residual life prediction method based on the RE-CF-EKF algorithm, which is characterized in that:
selecting a time domain characteristic representation capable of representing the whole service life trend of the bearing in the second step, and performing linear filtering on the time domain characteristic representation to obtain a filtered characteristic curve and a filtered error curve;
setting the size of a sliding window as m;
calculating the relative error RE (k) of F (k) and E (k) in each window, wherein the specific expression is as follows:
the RE curve is obtained and the RE curve,
selecting a threshold boundary line from an RE curve obtained from a current data set;
according to the threshold boundary, a degradation starting point is found and used as a prediction starting point TSP for subsequent prediction.
Generally, the degradation starting point is determined by the obtained RE curve and the selected threshold boundary, and when a certain value of the RE curve is greater than or equal to the threshold boundary, it is determined as the degradation point, for example: the predicted starting point TSP is the intersection of the RE curve and the relative error threshold boundary.
Further, the invention provides a rolling bearing residual life prediction method based on the RE-CF-EKF algorithm, which is characterized in that:
the sliding window size is X% of the known data length T;
the threshold boundary is the mean of the first A1% -A2% of the data of the relative error of the training set;
wherein X is a natural number;
a1 and A2 are selected from natural numbers, and A1 is less than A2.
The X, A1 and A2 can be adjusted according to the number of samples and actual needs.
Further, the invention provides a rolling bearing residual life prediction method based on the RE-CF-EKF algorithm, which is characterized in that:
in the third step, an improved algorithm CEEMDAN of empirical mode decomposition is used for reconstructing the variance characteristics and then carrying out noise reduction processing;
processing the variance characteristics after noise reduction according to an accumulation function to obtain an accumulation characteristic CF;
wherein the cumulative function is formulated as follows:
further, the invention provides a rolling bearing residual life prediction method based on the RE-CF-EKF algorithm, which is characterized in that:
the training set is life cycle data;
the test set takes the former part as known data, and the rest part is used for verifying the quality of the prediction result.
Further, the invention provides a rolling bearing residual life prediction method based on the RE-CF-EKF algorithm, which is characterized in that:
the fourth step comprises the following steps:
the EKF algorithm prediction test set known data state estimation value after last time TForming new model parameters;
obtaining bearing characteristic data corresponding to T + i (i is 1, 2, …, n) according to the prediction result;
when the characteristic curve of the bearing vibration data reaches a defined threshold value, the remaining service life RUL is the time interval from the time T to the bearing failure threshold point.
Further, the invention provides a rolling bearing residual life prediction method based on the RE-CF-EKF algorithm, which is characterized in that:
in the fourth step, the characteristic curve segment CF' (k) of the current data set more suitable for modeling prediction is obtained from the prediction starting point obtained in the second step and the accumulated characteristics obtained in the third step,k=TSP,…,T;
calculating the slope th of the last point of each characteristic curve segment of the training setiTaking the average value as a threshold value in the prediction of the test set:
the state space equation established is:
wherein x isk=[ak,bk,ck,dk];
k represents the number of cycles;
xkrepresenting state model parameters;
Ykan observed value representing a cumulative characteristic curve segment;
wkrepresenting state noise;
vkrepresenting observation noise;
fitting according to the characteristic curve segment of the training set to obtain initial parameters of a prediction model;
and inputting the initial model parameters and the characteristic curve segment of the current data of the test set into an EFK algorithm, and predicting to obtain a state value X (T) at the time T.
Further, the invention provides a rolling bearing residual life prediction method based on the RE-CF-EKF algorithm, which is characterized in that:
the specific flow of the EKF algorithm is as follows:
s1, considering the following state space equation:
Y(k)=HX(k)+V(k)
s2, prediction part:
s2-1. State prediction:
s2-2. Observation prediction:
Y(k|k-1)=HX(k|k-1)
s2-3. covariance prediction:
s3, updating:
s3-1, calculating a kalman gain:
K(k)=P(k|k-1)HT[HP(k|k-1)HT+R]-1
s3-2, status updating:
X(k)=X(k|k-1)+K(k)[Y(k)-Y(k|k-1)]
s3-3. covariance update:
P(k)=[In-K(k)H]P(k|k-1)。
further, the invention provides a rolling bearing residual life prediction method based on the RE-CF-EKF algorithm, which is characterized in that:
the calculation method of the measurement estimation value is as follows:
wherein Y (k) represents the input bearing cumulative characteristic curve segment CF' (k), corresponding to Y in the state space equationkObtaining the state value X at the time of T through an EKF algorithmkCalculating to obtain a measurement estimation value after T time according to a prediction model
When in useAnd when the threshold value is reached, stopping prediction, and calculating the predicted residual life value of the current bearing:
R=k-T
in which, for the characteristic curve segment CF' (k),k=TSP,…,Titeratively bringing the TSP time to t,TSP<t≤Tthe characteristic curve segment of the time is used as the input of the EKF, and the residual life is predicted.
The invention has the following functions and effects:
in the invention, a relative error curve is obtained by linearly filtering time domain characteristics and then sliding a window, and a prediction starting point is determined according to curve characteristics; then, obtaining more trend accumulated characteristics by selecting and processing time domain characteristics, and then constructing a state space equation based on the characteristic curve segment after the starting point is predicted; according to the established state space model, a state prediction value of the last point of the known data is obtained by using an extended Kalman filtering method, the subsequent trend of the characteristic curve is predicted according to the state prediction value and the prediction model, the prediction is stopped when the threshold value is reached, the corresponding residual service life is calculated, repeated tests are carried out by changing the size of the known data volume, and the accuracy and the robustness of the method are verified through the experimental results.
In addition, in the following embodiments, the present invention further evaluates two indexes, namely, a mean error and a root mean square error, of a plurality of remaining life values predicted by input samples with different data sizes and a true remaining life value as a test set, and compares the two indexes with other models and methods to verify the effectiveness of the method.
Drawings
Fig. 1 is a graph of a life cycle vibration signal of a Bearing used in the present embodiment, which is taken as an example of a training set Bearing 1-1.
FIG. 2 is a block diagram of the RE-CF-EKF algorithm used in this example.
FIG. 3 is a graph illustrating the effect of the method used in this embodiment on the Bearing1-3 data to determine the predicted starting point according to the relative error curve.
FIG. 4 is a comparison graph of the health index and the general variance characteristics constructed by the method used in the present embodiment on the Bearing1-3 data.
FIG. 5 is a comparison graph of the predicted effect of the method used in the present embodiment and other prediction models on Bearing1-3 data.
FIG. 6 is a table showing the comparison of the predicted effect indexes of the lower bearing data under the working conditions used in the present embodiment
Detailed Description
The technical solution of the present invention will be described in detail with reference to the accompanying drawings 1-6. Simulation was performed on Bearing data for condition one, but only a comparison graph of the simulation effect of Bearing1-3 Bearing data is presented here to verify the effectiveness of the algorithm.
As shown in fig. 1, the simulation experiment is performed based on data of the bearing in PHM2012 under the first working condition, and the validity of the algorithm is verified in a test set by simulation according to the given training set data.
As shown in fig. 2, the invention provides a method for predicting the residual life of a bearing based on an RE-CF-EKF algorithm, which comprises the following steps:
firstly, screening an obtained PHM2012 Bearing data set, selecting 7 bearings under a first working condition as an experimental data set, and defining according to the data set, taking Bearing1-1 and Bearing1-2 as training sets, and taking the remaining Bearing1-3 to Bearing1-7 as test sets. The training set is full life cycle data, the test set takes the former part as known data, and the rest part is used for verifying the quality of the prediction result.
Step two: firstly, time domain characteristics are obtained for an original signal, then simple linear filtering is carried out, finally, a relative error value (RE) between filtered data and filtering errors is calculated, a prediction starting point is determined based on an RE curve, and the specific flow is as follows:
and extracting time domain characteristics of the original vibration signals, wherein variance characteristics which can better show the service life trend of the bearing are selected.
And performing linear filtering on the square difference characteristic to obtain a filtered characteristic curve F (k) and a filtered error curve E (k).
The sliding window size is set to 5% of the known data length T.
The relative error RE (k) for windows F (k) and E (k) is calculated, i.e., the continuous curve in FIG. 3.
The mean of the first 10% -20% of the data of the relative error of the current data set is chosen as the threshold boundary, i.e. the dashed line in fig. 3.
The degradation start point is determined from the first 20% data point according to the threshold boundary and used as a prediction start point for a subsequent prediction, i.e., the TSP shown in fig. 3.
Step three: the selected time domain features, namely the continuous curves in fig. 4, are subjected to noise reduction processing, and then an accumulation function is used to obtain an accumulation feature (CF) with a smoother trend, wherein the specific flow is as follows:
and (3) reconstructing the variance characteristics by using an improved algorithm CEEMDAN of empirical mode decomposition and then performing noise reduction.
The variance feature after noise reduction is processed according to an accumulation function to obtain an accumulation feature (CF), i.e. a dotted line in fig. 4, and the accumulation function formula is as follows:
step four: and (4) combining the second step and the third step to obtain a characteristic curve more suitable for modeling prediction, and establishing a state space model. Obtaining initial parameters of a model by a characteristic curve of a training set, and predicting the residual life of a bearing by combining an EKF algorithm, wherein the specific process comprises the following steps:
obtaining a characteristic curve segment CF' (k) of the current data set which is more suitable for modeling and prediction by the TSP obtained in the second step and the CF (k) obtained in the third step,k=TSP,…,T
calculating the slope th of the last point of each characteristic curve segment of the training setiTaking the average value as a threshold value in the prediction of the test set:
the state space equation established is:
wherein x isk=[ak,bk,ck,dk]K denotes the number of cycles, xkRepresenting a state model parameter, YkRepresents an observed value of the cumulative characteristic curve segment, and wkAnd vkRepresenting state noise and observation noise, respectively.
And fitting the characteristic curve segments of the training set to obtain initial parameters of the prediction model. Inputting the initial model parameters and the characteristic curve segment of the current data of the test set into an EFK algorithm, and predicting to obtain a state value X (T) at the time T:
(1) consider the following state space equation:
Y(k)=HX(k)+V(k)
(2) and a prediction part:
state prediction:
observation and prediction:
Y(k|k-1)=HX(k|k-1)
and thirdly, covariance prediction:
(3) an updating part:
computing a kalman gain:
K(k)=P(k|k-1)HT[HP(k|k-1)HT+R]-1
state updating:
X(k)=X(k|k-1)+K(k)[Y(k)-Y(k|k-1)]
updating the covariance:
P(k)=[In-K(k)H]P(k|k-1)
calculating a measurement estimation value:
where Y (k) represents the input bearing cumulative characteristic segment CF' (k), corresponding to Y in the state space equationkObtaining the state value X at the time of T through an EKF algorithmkThen, according to the prediction model, the measurement estimation value after T time is calculated and obtained
When in useAnd when the threshold value is reached, stopping prediction, and calculating the predicted residual life value of the current bearing:
R=k-T
for the characteristic curve segment CF' (k),k=TSP,…,Titeratively bringing the TSP time to t,TSP<t≤Tthe characteristic curve segment CF' (k) of the time is used as an input for the EKF and the remaining life is predicted.
And evaluating two indexes, namely mean error and root mean square error, of a plurality of residual life values predicted by input samples with different data sizes and real residual life values serving as a test set, comparing the indexes with other models and methods, and verifying the effectiveness of the method.
And evaluating two indexes, namely mean error and root mean square error, of a plurality of residual life values predicted by input samples with different data sizes and real residual life values serving as a test set, comparing the indexes with other models and methods, and verifying the effectiveness of the method.
FIG. 5 illustrates Bearing data of Bearing1-3, and compares the bi-exponential model used in the method with the Weibull model and Paris model, and sets a confidence interval of + -20%. As can be seen from the comparison result of the prediction result in FIG. 5 and the index in FIG. 6, the proposed RE-CF-EKF algorithm has better residual life prediction effect and prediction robustness.
Aiming at the problem of residual life prediction of a rolling bearing, the invention provides an RE-CF-EKF algorithm, a simulation experiment is realized based on PHM2012 bearing data, and the effectiveness of the RE-CF-EKF algorithm is proved through the experiment.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (9)
1. A rolling bearing residual life prediction method based on an RE-CF-EKF algorithm is characterized by comprising the following steps:
the method comprises the following steps: screening the obtained bearing vibration data set, selecting a plurality of bearings under the same working condition as an experimental data set, and grouping the experimental data set with the bearings under the same working condition into a training set and a test set;
step two: linear filtering is carried out after the time domain characteristics of the original signal are taken, a relative error value RE between filtered data and a filtering error is calculated, and a prediction starting point TSP is determined based on an RE curve;
step three: denoising the selected time domain characteristics, and obtaining an accumulated characteristic CF with smoother trend by an accumulated function;
step four: and combining the second step and the third step to obtain a characteristic curve section suitable for modeling prediction, establishing a state space model, obtaining model initial parameters from the characteristic curve section of the training set, and then combining an EKF algorithm to predict the residual life of the bearing.
2. The rolling bearing residual life prediction method based on the RE-CF-EKF algorithm as claimed in claim 1, characterized in that:
selecting a time domain characteristic representation capable of representing the whole service life trend of the bearing in the second step, and performing linear filtering on the time domain characteristic representation to obtain a filtered characteristic curve and a filtered error curve;
setting the size of a sliding window as m;
calculating the relative error RE (k) of F (k) and E (k) in each window, wherein the specific expression is as follows:
the RE curve is obtained and the RE curve,
selecting a threshold boundary line from an RE curve obtained from a current data set;
according to the threshold boundary, a degradation starting point is found and used as a prediction starting point TSP for subsequent prediction.
3. The rolling bearing residual life prediction method based on the RE-CF-EKF algorithm as claimed in claim 2, characterized in that:
the sliding window size is X% of the known data length T;
the threshold boundary is the mean of the first A1% -A2% of the data of the relative error of the training set;
wherein X is a natural number;
a1 and A2 are selected from natural numbers, and A1 is less than A2.
4. The rolling bearing residual life prediction method based on the RE-CF-EKF algorithm as claimed in claim 1, characterized in that:
in the third step, an improved algorithm CEEMDAN of empirical mode decomposition is used for reconstructing the variance characteristics and then carrying out noise reduction processing;
processing the variance feature V subjected to noise reduction according to an accumulation function to obtain an accumulation feature CF;
wherein the cumulative function is formulated as follows:
5. the rolling bearing residual life prediction method based on the RE-CF-EKF algorithm as claimed in claim 1, characterized in that:
the training set is life cycle data;
the test set takes the former part as known data, and the rest part is used for verifying the quality of the prediction result.
6. The rolling bearing residual life prediction method based on the RE-CF-EKF algorithm as claimed in claim 5, characterized in that:
the fourth step comprises the following steps:
the EKF algorithm predicts the state estimation value after the last time T of the known data of the test setForming new model parameters;
obtaining bearing characteristic data corresponding to T + i (i is 1, 2, …, n) according to the prediction result;
when the characteristic curve of the bearing vibration data reaches a defined threshold value, the remaining service life RUL is the time interval from the time T to the bearing failure threshold point.
7. The rolling bearing residual life prediction method based on the RE-CF-EKF algorithm as claimed in claim 5, characterized in that:
in the fourth step, the characteristic curve segment CF' (k) of the current data set more suitable for modeling prediction is obtained from the prediction starting point obtained in the second step and the accumulated characteristics obtained in the third step,k=TSP,…,T;
calculating the slope th of the last point of each characteristic curve segment of the training setiTaking the average value as a threshold value in the prediction of the test set:
the state space equation established is:
wherein x isk=[ak,bk,ck,dk];
k represents the number of cycles;
xkrepresenting state model parameters;
Ykan observed value representing a cumulative characteristic curve segment;
wkrepresenting state noise;
vkrepresenting observation noise;
fitting according to the characteristic curve segment of the training set to obtain initial parameters of a prediction model;
and inputting the initial model parameters and the characteristic curve segment of the current data of the test set into an EFK algorithm, and predicting to obtain a state value X (T) at the time T.
8. The rolling bearing residual life prediction method based on the RE-CF-EKF algorithm as claimed in claim 7, wherein:
the specific flow of the EKF algorithm is as follows:
s1, considering the following state space equation:
Y(k)=HX(k)+V(k)
s2, prediction part:
s2-1. State prediction:
s2-2. Observation prediction:
Y(k|k-1)=HX(k|k-1)
s2-3. covariance prediction:
s3, updating:
s3-1, calculating a kalman gain:
K(k)=P(k|k-1)HT[HP(k|k-1)HT+R]-1
s3-2, status updating:
X(k)=X(k|k-1)+K(k)[Y(k)-Y(k|k-1)]
s3-3. covariance update:
P(k)=[In-K(k)H]P(k|k-1)。
9. the rolling bearing residual life prediction method based on the RE-CF-EKF algorithm as claimed in claim 8, characterized in that:
the calculation method of the measurement estimation value is as follows:
wherein Y (k) represents the input bearing cumulative characteristic curve segment CF' (k), corresponding to Y in the state space equationkObtaining the state value X at the time of T through an EKF algorithmkCalculating to obtain a measurement estimation value after T time according to a prediction model
When in useAnd when the threshold value is reached, stopping prediction, and calculating the predicted residual life value of the current bearing:
R=k-T
in which, for the characteristic curve segment CF' (k),k=TSP,…,Titeratively bringing the TSP time to t,TSP<t≤Tthe characteristic curve segment of the time is used as the input of the EKF, and the residual life is predicted.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010968660.4A CN112100919A (en) | 2020-09-15 | 2020-09-15 | Rolling bearing residual life prediction method based on RE-CF-EKF algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010968660.4A CN112100919A (en) | 2020-09-15 | 2020-09-15 | Rolling bearing residual life prediction method based on RE-CF-EKF algorithm |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112100919A true CN112100919A (en) | 2020-12-18 |
Family
ID=73760141
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010968660.4A Pending CN112100919A (en) | 2020-09-15 | 2020-09-15 | Rolling bearing residual life prediction method based on RE-CF-EKF algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112100919A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112762100A (en) * | 2021-01-14 | 2021-05-07 | 哈尔滨理工大学 | Bearing full-life-cycle monitoring method based on digital twinning |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
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 |
CN106021719A (en) * | 2016-05-19 | 2016-10-12 | 浙江工业大学 | Unscented Kalman filter algorithm-based method for predicting residual life of bearing |
CN106909756A (en) * | 2017-03-29 | 2017-06-30 | 电子科技大学 | A kind of rolling bearing method for predicting residual useful life |
CN109977491A (en) * | 2019-03-06 | 2019-07-05 | 北京航空航天大学 | A kind of impact injury can restore under the conditions of degeneration modeling and life-span prediction method |
GB201912597D0 (en) * | 2019-09-02 | 2019-10-16 | Siemens Plc | Method and system for predicting battery degredation |
KR20200049295A (en) * | 2018-10-31 | 2020-05-08 | 한국전력공사 | A method to predict health index transition and residual life for turbomachinery |
-
2020
- 2020-09-15 CN CN202010968660.4A patent/CN112100919A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
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 |
CN106021719A (en) * | 2016-05-19 | 2016-10-12 | 浙江工业大学 | Unscented Kalman filter algorithm-based method for predicting residual life of bearing |
CN106909756A (en) * | 2017-03-29 | 2017-06-30 | 电子科技大学 | A kind of rolling bearing method for predicting residual useful life |
KR20200049295A (en) * | 2018-10-31 | 2020-05-08 | 한국전력공사 | A method to predict health index transition and residual life for turbomachinery |
CN109977491A (en) * | 2019-03-06 | 2019-07-05 | 北京航空航天大学 | A kind of impact injury can restore under the conditions of degeneration modeling and life-span prediction method |
GB201912597D0 (en) * | 2019-09-02 | 2019-10-16 | Siemens Plc | Method and system for predicting battery degredation |
Non-Patent Citations (4)
Title |
---|
CAIXIA GAO等: "Method for Determining Starting Point of Rolling Bearing Life Prediction Based on Linear Regression", 《ELECTRONICS》, vol. 8, no. 9, 22 August 2019 (2019-08-22), pages 1 - 10 * |
SALMA HARRATH等: "A New Adaptive Prognostic Strategy Based on Online Future Evaluation and Extended Kalman Filtering", 《2019 6TH INTERNATIONAL CONFERENCE ON CONTROL, DECISION AND INFORMATION TECHNOLOGIES (CODIT 2019)》, 26 April 2019 (2019-04-26), pages 2033 - 2038, XP033610536, DOI: 10.1109/CoDIT.2019.8820471 * |
SEN ZHAO等: "A recurrent neural network approach for remaining useful life prediction utilizing a novel trend features construction method", 《MEASUREMENT》, vol. 146, 17 June 2019 (2019-06-17), pages 279 - 288, XP085761068, DOI: 10.1016/j.measurement.2019.06.004 * |
高彩霞等: "线性回归与EEMD的滚动轴承剩余寿命预测", 《机械科学与技术》, vol. 38, no. 10, 31 October 2019 (2019-10-31), pages 1589 - 1597 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112762100A (en) * | 2021-01-14 | 2021-05-07 | 哈尔滨理工大学 | Bearing full-life-cycle monitoring method based on digital twinning |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110018670B (en) | Industrial process abnormal working condition prediction method based on dynamic association rule mining | |
CN111709465B (en) | Intelligent identification method for rough difference of dam safety monitoring data | |
CN111639430A (en) | Digital twin driven natural gas pipeline leakage identification system | |
CN105205288A (en) | Mode evolution-based forecasting method for satellite long-term on-orbit running state | |
CN108304348B (en) | Bearing residual life prediction method based on binary wiener process | |
CN109271319B (en) | Software fault prediction method based on panel data analysis | |
EP1960853A1 (en) | Evaluating anomaly for one-class classifiers in machine condition monitoring | |
CN116226646B (en) | Method, system, equipment and medium for predicting health state and residual life of bearing | |
CN110596506A (en) | Converter fault diagnosis method based on time convolution network | |
CN113642754A (en) | Complex industrial process fault prediction method based on RF noise reduction self-coding information reconstruction and time convolution network | |
CN113255835A (en) | Hydropower station pump equipment anomaly detection method | |
CN112683535B (en) | Bearing life prediction method based on multi-stage wiener process | |
CN114167838B (en) | Multi-scale health assessment and fault prediction method for servo system | |
Qin et al. | Remaining useful life prediction for rotating machinery based on optimal degradation indicator | |
CN113987908A (en) | Natural gas pipe network leakage early warning method based on machine learning method | |
CN114912640A (en) | Method and system for detecting abnormal mode of generator set based on deep learning | |
CN113158471A (en) | Degradation equipment residual life prediction method considering measurement uncertainty | |
CN114897103A (en) | Industrial process fault diagnosis method based on neighbor component loss optimization multi-scale convolutional neural network | |
CN112100919A (en) | Rolling bearing residual life prediction method based on RE-CF-EKF algorithm | |
CN111338310B (en) | Industrial process steady-state working condition identification and classification method | |
CN111241629B (en) | Intelligent prediction method for performance change trend of hydraulic pump of airplane based on data driving | |
CN110956112B (en) | Novel high-reliability slewing bearing service life assessment method | |
CN103400201A (en) | Method for solving state estimation problem taking maximum normal rate of measurement point as target | |
CN114112390B (en) | Nonlinear complex system early fault diagnosis method | |
CN115098962A (en) | Method for predicting residual life of mechanical equipment in degradation state based on hidden half Markov model |
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 |