CN115046764B - Early fault diagnosis method for rolling bearing - Google Patents

Early fault diagnosis method for rolling bearing Download PDF

Info

Publication number
CN115046764B
CN115046764B CN202210482268.8A CN202210482268A CN115046764B CN 115046764 B CN115046764 B CN 115046764B CN 202210482268 A CN202210482268 A CN 202210482268A CN 115046764 B CN115046764 B CN 115046764B
Authority
CN
China
Prior art keywords
frequency
value
rolling bearing
characteristic
calculating
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
Application number
CN202210482268.8A
Other languages
Chinese (zh)
Other versions
CN115046764A (en
Inventor
孟宗
藩红苗
樊凤杰
李继猛
曹利宵
葛少普
陈昌
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Yanshan University
Original Assignee
Yanshan University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Yanshan University filed Critical Yanshan University
Priority to CN202210482268.8A priority Critical patent/CN115046764B/en
Publication of CN115046764A publication Critical patent/CN115046764A/en
Application granted granted Critical
Publication of CN115046764B publication Critical patent/CN115046764B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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

Landscapes

  • Physics & Mathematics (AREA)
  • Acoustics & Sound (AREA)
  • General Physics & Mathematics (AREA)
  • Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
  • Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)

Abstract

The invention discloses a rolling bearing early fault diagnosis method, which belongs to the technical field of mechanical state monitoring, and aims to overcome the defect that the traditional evaluation index is insensitive to early faults. And then, by utilizing a continuous alarm frequency triggering mechanism, the early failure of the rolling bearing can be accurately and reliably monitored only by using a smaller triggering frequency. The research result of the invention in the aspect of early fault diagnosis of the rolling bearing lays a foundation for realizing the residual life prediction of the rolling bearing.

Description

Early fault diagnosis method for rolling bearing
Technical Field
The invention belongs to the technical field of mechanical state monitoring, and particularly relates to a method for diagnosing early faults of a rolling bearing.
Background
The rolling bearing is used as a key component in a transmission system of mechanical equipment, and is widely applied to the mechanical equipment. The working environment of the rolling bearing is often very complex, and the rolling bearing is usually continuously operated under the conditions of high rotating speed and high load, so that faults are easy to occur. When it is in a fault operation state, it has a serious influence on the health condition of the rotating mechanical equipment, and even finally leads to the failure of the equipment. In view of this, in order to ensure the safety and reliability of the equipment and avoid the occurrence of catastrophic accidents, the working state of the bearing must be strictly monitored, and preferably real-time online detection is performed to prevent accidents caused by rolling bearing failures.
Early fault monitoring and diagnosis of the rolling bearing are hot spots in research in the current industrial field, particularly, with the development of intelligent technology, the automation degree of mechanical equipment is higher and higher, and the rolling bearing is a key component of the mechanical equipment, so that higher requirements on the running state monitoring of the rolling bearing are also provided. The fault characteristics commonly used by engineering technicians at present are time domain characteristics, frequency domain characteristics and time-frequency domain characteristics. When the early fault occurs in the running of the rolling bearing, the fluctuation of the vibration signal is weak, and the difference between the extracted characteristic parameter value of the early fault and the characteristic parameter value in the stationary period is not large. For different bearings and different working conditions of the same type of bearings, the performance changes of the characteristics are large, and no general indicator effectively reflects the running state of the bearings in different states. If these characteristics of the signal are directly utilized, the signal is insensitive to early failure of the rolling bearing, so that detection omission of early failure is caused, and the detection is only carried out when the rolling bearing has obvious failure. Most of the characteristics are isolated and not connected with each other, and the multi-dimensional characteristic information is slightly redundant. A continuous trigger mechanism adopted in engineering is used for judging early faults of machinery, namely when M continuous indexes exceed a threshold value, the early faults are indicated. However, the value of M is generally large, and the determination cannot be performed with a small value of M. The above-described disadvantage may greatly affect the ability to detect an early failure time, and the discrimination accuracy thereof may be degraded.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide an accurate and rapid early fault diagnosis method for a rolling bearing.
In order to achieve the purpose, the technical scheme of the invention comprises the following steps:
1) And acquiring initial vibration data S of the whole life cycle of the rolling bearing. Selecting p pieces of history data acquired initially, extracting A pieces of characteristics to form an original characteristic index set F = (F) 1 ,f 2 ,...,f i ) I =1, 2.., a, wherein f i ＝(f i1 ,f i2 ,...,f ip )。
2) Carrying out Gaussian denoising algorithm smoothing treatment on each feature in the original feature set F to obtain F '= (F' 1 ,f' 2 ,...,f' i )。
Then, each feature in the feature set F' is subjected to evolution cumulative summation technology conversion processing to obtain CF = (CF) 1 ,cf 2 ,...,cf i )。
3) Calculating Pearson correlation coefficients of A-1 features of the evolution cumulative sum root-mean-square and other evolution cumulative sum transformation in the feature set, setting a preset threshold q, and reserving B features larger than the threshold q to obtain a new feature index set CF' = (CF =) which is 1 ',cf 2 ',...,cf' B )。
4) And (4) carrying out principal component analysis calculation on the new characteristic index set CF', and selecting a difference array of the first principal component array as the health index HI. Calculating the mean value of the health index by using p-1 initial historical data of the health index
Figure GDA0004094976910000021
And standard deviation sigma, calculating an initial threshold value>
Figure GDA0004094976910000022
Calculating A characteristics of the new vibration signal collected in real time, adding the A characteristics into F', still utilizing the characteristic set selected from the steps 1) to 3), and having p +1 values in each characteristic. And then, performing principal component analysis and calculation on the characteristics, taking the first principal component of the characteristics, and calculating the difference number of the first principal component as a new health index HI'.
If there are consecutive 3 values of HI' m 、HI' m+1 And HI' m+2 Exceeding the threshold T indicates a time T m Early failure of the rolling bearing occurs. And if no continuous 3 values exceed the threshold value T, reading the next vibration signal, calculating A characteristics of the new vibration signal, adding the A characteristics to the last F', reusing the characteristic sets selected in the steps 1-3, then carrying out principal component analysis calculation on the characteristics, and calculating the health index until the continuous 3 health index values exceed the threshold value T.
Further, the features in step 1) include 14 time-domain features and 4 frequency-domain features.
The 14 time domain features include 10 dimensional time domain features, which are respectively root mean square, maximum value, minimum value, average value, peak-to-peak value, absolute average value, variance, standard deviation, kurtosis and skewness, and the specific expression is as follows:
root mean square value
Figure GDA0004094976910000031
Maximum value F 2 ＝max(x(i))；
Minimum value F 3 = min (x (i)); mean value
Figure GDA0004094976910000032
Peak to peak value F 5 = max (x (i)) -min (x (i)); absolute average value
Figure GDA0004094976910000033
Variance (variance)
Figure GDA0004094976910000034
Standard deviation is greater or less>
Figure GDA0004094976910000035
Kurtosis
Figure GDA0004094976910000036
Deflection->
Figure GDA0004094976910000037
The 4 dimensionless time domain characteristics are a form factor, a peak factor, a pulse factor and a margin factor respectively, and the specific expression is as follows:
form factor
Figure GDA0004094976910000038
Peak factor->
Figure GDA0004094976910000039
Pulse factor
Figure GDA00040949769100000310
Margin factor>
Figure GDA00040949769100000311
Among these characteristic parameters are those that are sensitive to early bearing failure, such as kurtosis, peaks, margin factors, and impulse factors, but are primarily directed to impact-type failures. There are also characteristic parameters that change more steadily during the overall degradation process, such as root mean square values, which change correspondingly as the degree of bearing wear increases. Therefore, these indexes are generally used in combination in consideration of both sensitivity and stability.
The 4 frequency domain features include the mean frequency, center of gravity frequency, root mean square frequency, and frequency standard deviation. In order to obtain the frequency domain characteristics of the vibration signal, the time domain signal may be subjected to a fourier transform FFT to obtain a frequency domain signal. The specific expression is as follows:
mean frequency
Figure GDA0004094976910000041
Center of gravity frequency>
Figure GDA0004094976910000042
Root mean square frequency
Figure GDA0004094976910000043
Frequency standard deviation>
Figure GDA0004094976910000044
Wherein f is k And s (k) is the corresponding frequency amplitude of the k-th spectral line.
k＝1,2,...,N fft ，N fft Is the length of the spectral signal.
Frequency characteristic F 15 Reflecting the magnitude of the frequency vibration energy, F 16 ～F 17 Reflecting the change of the position of the main frequency band, F 18 Reflecting the degree of spectral dispersion.
Further, in the step 2), smoothing is performed by using a gaussian denoising algorithm, that is, a gaussian weighted moving average filter, and the process is as follows:
let the sliding window length be n. When n is even or odd:
Figure GDA0004094976910000045
where G is a weight vector satisfying a Gaussian distribution, f i ' is the feature of the ith feature after the smoothing of the Gaussian weighted moving average filter. The expression of the gaussian weight vector G is:
Figure GDA0004094976910000051
where A is the normalized coefficient, a is the width of the Gaussian denoising window,
Figure GDA0004094976910000052
and the formula of the evolution cumulative sum algorithm is as follows:
Figure GDA0004094976910000053
wherein
Figure GDA0004094976910000054
The method is characterized in that the ith characteristic of the bearing in the initial operation period is taken as an average value;
further, the formula of the pearson correlation coefficient in step 3) is as follows:
Figure GDA0004094976910000055
/>
where k is the length of the feature x (i),
Figure GDA0004094976910000056
and &>
Figure GDA0004094976910000057
Are the mean values of x (i) and y (i), respectively. The Pearson correlation coefficient has a value range of [ -1,1]. The closer the absolute value of the correlation coefficient is to 1, the higher the correlation degree of x and y is; the closer the absolute value of the correlation coefficient is to 0, the lower the correlation of x with y.
The invention has the advantages that: firstly, smoothing noisy signals by using a Gaussian weighted moving average filter, secondly, transforming the characteristics by using an evolution accumulation addition and transformation technology, overcoming the defect that the traditional evaluation index is insensitive to early faults, secondly, converting multidimensional characteristics into single-dimensional characteristics by using a principal component analysis method in order to fully fuse characteristic information of a time domain and a frequency domain, and taking a difference array as a health index, wherein the change of early degradation of the rolling bearing can be represented to the maximum extent. And finally, a continuous alarm frequency triggering mechanism is utilized, and the early failure of the rolling bearing can be accurately and reliably monitored only by using a small triggering frequency, so that the interference of external factors such as noise and the like is overcome, and the early failure can be rapidly judged.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a graph of vibration signals monitored by two sets of test bearings according to an embodiment;
FIG. 3 is a diagram showing the result of the judgment of the bearing early failure occurrence time by the method of the present invention in the bearing 1 according to the embodiment of the present invention;
FIG. 4 is a diagram showing the result of the judgment of the bearing early failure occurrence time by the method of the present invention in the bearing 2 of the embodiment.
Detailed description of the invention
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings.
The general flow chart of the method for diagnosing the early failure of the rolling bearing is shown in figure 1, and the technical scheme and the implementation steps of the method are as follows:
step 1, acquiring initial vibration data S of a whole life cycle of a rolling bearing, and selecting initial p data to perform feature extraction;
step 2, carrying out Gaussian weighted moving average filtering processing and evolution accumulation addition technology conversion processing on the features;
step 3, selecting characteristics by utilizing a Pearson correlation coefficient;
and 4, constructing a new health index by utilizing a principal component analysis technology, calculating a threshold value, and monitoring early faults of the rolling bearing in real time by utilizing a continuous alarm frequency triggering mechanism.
The respective steps are explained in detail one by one below.
Step 1: data are from PHM 2012PRONOSTIA Rolling bearing accelerated life test bench. In the full-life accelerated life experiment, the sampling frequency of the acceleration sensor is 25.6kHz, the sampling time is 0.1s, the sampling interval is 10s, and 2560 points are sampled each time. The shaft speed was 1800rmp and the bearing radial load was 4000N.
We use two degradation trends for the bearing embodiment, one is slow degradation and one is sudden degradation, as shown in fig. 2 for bearing 1 and bearing 2 respectively. As can be seen from the figure, when the bearing initially runs, the vibration signal is stable and is in a normal running state; over time, the amplitude of the vibration signal is gradually increased or suddenly increased, and at the moment, the rolling bearing is in a failure state and is in a degradation state.
Selecting 150 pieces of initially acquired historical data, extracting 18 pieces of features, and forming an original feature index set F = (F) 1 ,f 2 ,...,f i ) I =1, 2.., 18, wherein f i ＝(f i1 ,f i2 ,...,f i150 )。
The 18 features include 14 time domain features and 4 frequency domain features.
The 14 time domain features include 10 dimensional time domain features, which are respectively root mean square, maximum value, minimum value, average value, peak-to-peak value, absolute average value, variance, standard deviation, kurtosis and bias, and the specific expression is as follows:
root mean square value
Figure GDA0004094976910000071
Maximum value F 2 ＝max(x(i))；
Minimum value F 3 = min (x (i)); mean value
Figure GDA0004094976910000072
Peak to peak value F 5 = max (x (i)) -min (x (i)); absolute mean value
Figure GDA0004094976910000073
Variance (variance)
Figure GDA0004094976910000074
Standard deviation->
Figure GDA0004094976910000075
Kurtosis
Figure GDA0004094976910000076
Deflection degree->
Figure GDA0004094976910000077
The 4 dimensionless time domain characteristics are a form factor, a peak factor, a pulse factor and a margin factor respectively, and the specific expression is as follows:
form factor
Figure GDA0004094976910000078
Peak factor->
Figure GDA0004094976910000079
Pulse factor
Figure GDA00040949769100000710
Margin factor>
Figure GDA00040949769100000711
Among these characteristic parameters are those that are sensitive to early bearing failure, such as kurtosis, peaks, margin factors, and impulse factors, but are primarily directed to impact-type failures. There are also characteristic parameters that change more steadily during the overall degradation process, such as root mean square value, which changes correspondingly as the degree of bearing wear increases. Therefore, these indexes are generally used in combination from the viewpoint of both sensitivity and stability.
The 4 frequency domain features include the mean frequency, center of gravity frequency, root mean square frequency, and frequency standard deviation. In order to obtain the frequency domain characteristics of the vibration signal, the time domain signal may be subjected to a fourier transform FFT to obtain a frequency domain signal. The specific expression is as follows:
mean frequency
Figure GDA0004094976910000081
Center of gravity frequency->
Figure GDA0004094976910000082
Root mean square frequency
Figure GDA0004094976910000083
Frequency standard deviation>
Figure GDA0004094976910000084
Wherein f is k And s (k) is the corresponding frequency amplitude of the k-th spectral line.
k＝1,2,...,N fft ，N fft Is the length of the spectral signal.
Frequency characteristic F 15 Reflecting the magnitude of the frequency vibration energy, F16-F 17 Reflecting the change of the position of the main band, F 18 Reflecting the degree of spectral dispersion.
And 2, step: performing Gaussian denoising algorithm smoothing processing, namely Gaussian weighted moving average filtering, on each feature in the original feature set F to obtain F '= (F' 1 ,f 2 ',...,f i ') the procedure is as follows:
let the sliding window length be 45.
Figure GDA0004094976910000085
Where G is a weight vector satisfying a Gaussian distribution, f i ' is the feature of the ith feature after the smoothing of the Gaussian weighted moving average filter. The expression of the gaussian weight vector G is:
Figure GDA0004094976910000086
wherein A =1, a =45,
Figure GDA0004094976910000087
then, each feature in the feature set F' is subjected to evolution cumulative summation technology conversion processing to obtain CF = (CF) 1 ,cf 2 ,...,cf i )。
The formula of the evolution cumulative summation algorithm is as follows:
Figure GDA0004094976910000091
wherein
Figure GDA0004094976910000092
The average value of the ith characteristic of the bearing in the initial operation period is taken;
and step 3: calculating Pearson correlation coefficients of A-1 features of the root mean square cumulative sum and other evolution cumulative sum transformation in the feature set, setting a preset threshold value of 0.98, and reserving B features larger than the threshold value of 0.98 to obtain a new feature index set CF' = (CF) 1 ',cf 2 ',...,cf' B )。
The formula for the pearson correlation coefficient is as follows:
Figure GDA0004094976910000093
where k is the length of the feature x (i),
Figure GDA0004094976910000094
and &>
Figure GDA0004094976910000095
Are the mean values of x (i) and y (i), respectively. The Pearson correlation coefficient has a value range of [ -1,1]. The closer the absolute value of the correlation coefficient is to 1, the higher the correlation degree of x and y is; the closer the absolute value of the correlation coefficient is to 0, the lower the correlation of x with y.
And 4, step 4: and (4) carrying out principal component analysis calculation on the new characteristic index set CF', and selecting a difference array of the first principal component array as the health index HI. Calculating to obtain the average value of the health index by using 149 initial historical data of the health index
Figure GDA0004094976910000096
And standard deviation sigma, calculating an initial threshold value->
Figure GDA0004094976910000097
Calculating 18 characteristics of the new vibration signal acquired in real time, adding the characteristics into F', and still utilizing the characteristic set selected from the steps 1) to 3), wherein each characteristic has 151 values. And then, performing principal component analysis and calculation on the characteristics, taking the first principal component of the characteristics, and calculating the difference number of the first principal component as a new health index HI'.
If there are consecutive 3 values of HI' m 、HI' m+1 And HI' m+2 Exceeding the threshold T indicates a time T m Early failure of the rolling bearing occurs. And if no continuous 3 values exceed the threshold value T, reading the next vibration signal, calculating 18 characteristics of the new vibration signal, adding the characteristics into the last F', reusing the characteristic sets selected in the steps 1-3, then carrying out principal component analysis calculation on the characteristics, and calculating the health index until the continuous 3 health index values exceed the threshold value T.
The method for constructing the health index provided by the method is adopted, the method for determining the early fault is adopted to judge the early fault of the bearing, and the monitoring result is shown in figure 3 and figure 4. As can be seen from the figure, the health indicator of the bearing varies greatly in the early stage and then gradually increases after returning to a certain value. Since 150 values (namely 1500 seconds) of the initial vibration signal of the rolling bearing are used as the basis for the subsequent data processing, the first 1500 seconds are removed when the fault time of the bearing is monitored in real time, and the fault time of the bearing is monitored only after the 151 th value is added. Even if extreme health indicator values suddenly change in the middle of a slow increase, we can judge that this is an impulse signal caused by environmental noise, rather than a true fault signal. Until the bearing 1 to 14040 seconds and the bearing 2 to 24110 seconds, the continuous 3 values in the health index of the rolling bearing exceed the respective threshold values, that is, the rolling bearing has an early failure.
Two specific cases are used in the application to show the principle and the detection effect of the invention, and the cases can help to understand the idea and the using method of the invention. These should not be construed as limitations of the invention, but all modifications made based on the idea of the invention should be within the scope of protection of the invention.

Claims (4)

1. A method for diagnosing early failure of a rolling bearing is characterized by comprising the following steps:
step 1, acquiring initial vibration data S of a rolling bearing in a whole life cycle, selecting p pieces of initially acquired historical data, extracting A pieces of characteristics, and forming an original characteristic index set F = (F) 1 ,f 2 ,...,f i ) I =1, 2.., a, wherein f i ＝(f i1 ,f i2 ,...,f ip )；
Step 2, performing Gaussian weighted moving average filtering processing and evolution accumulation summation technology transformation processing on the features, and performing Gaussian denoising algorithm smoothing processing on each feature in the original feature set F to obtain F '= (F' 1 ,f 2 ',...,f i ') to a host; then, each feature in the feature set F' is subjected to evolution cumulative summation technology conversion processing to obtain CF = (CF) 1 ,cf 2 ,...,cf i )；
Step 3, selecting by using Pearson correlation coefficientSelecting characteristics, calculating Pearson correlation coefficients of A-1 characteristics of the root mean square cumulative sum of the evolution in the characteristic set and the cumulative sum of other evolution, setting a preset threshold q, and reserving B characteristics larger than the threshold q to obtain a new characteristic index set CF' = (CF) 1 ',cf 2 ',...,cf B ')；
Step 4, constructing a new health index by utilizing a principal component analysis technology, calculating a threshold value, and monitoring early faults of the rolling bearing in real time by utilizing a continuous alarm frequency triggering mechanism; performing principal component analysis calculation on the new characteristic index set CF', and selecting a differential array of the first principal component array as a health index HI; calculating to obtain the average value of the health index by utilizing p-1 initial historical data of the health index
Figure QLYQS_1
And standard deviation sigma, calculating an initial threshold value->
Figure QLYQS_2
Calculating A characteristics of the new vibration signal collected in real time, adding the characteristics into F', using the characteristic set selected in the steps 1-3, wherein each characteristic has p +1 value; then, performing principal component analysis and calculation on the characteristics, taking the first principal component, and calculating the difference number of the first principal component as a new health index HI';
if there are consecutive 3 values of HI' m 、HI' m+1 And HI' m+2 Exceeding the threshold T indicates a time T m Early failure of the rolling bearing occurs; and if no continuous 3 values exceed the threshold value T, reading the next vibration signal, calculating A characteristics of the new vibration signal, adding the A characteristics to the last F', reusing the characteristic sets selected in the steps 1-3, then carrying out principal component analysis calculation on the characteristics, and calculating the health index until the continuous 3 health index values exceed the threshold value T.
2. The early failure diagnosis method for a rolling bearing according to claim 1, characterized in that: the characteristics in the step 1 comprise 14 time domain characteristics and 4 frequency domain characteristics; the 14 time domain features include 10 dimensional time domain features, which are respectively root mean square, maximum value, minimum value, average value, peak-to-peak value, absolute average value, variance, standard deviation, kurtosis and skewness, and the specific expression is as follows:
root mean square value
Figure QLYQS_3
Maximum value F 2 = max (x (i)); minimum value F 3 = min (x (i)); mean value->
Figure QLYQS_4
Peak to peak value F 5 = max (x (i)) -min (x (i)); absolute mean value
Figure QLYQS_5
Variance (variance)
Figure QLYQS_6
Standard deviation is greater or less>
Figure QLYQS_7
/>
Kurtosis
Figure QLYQS_8
Deflection degree->
Figure QLYQS_9
The 4 dimensionless time domain characteristics are a form factor, a peak factor, a pulse factor and a margin factor respectively, and the specific expression is as follows:
form factor
Figure QLYQS_10
Peak factor->
Figure QLYQS_11
Pulse factor
Figure QLYQS_12
Margin factor->
Figure QLYQS_13
The 4 frequency domain characteristics comprise average frequency, barycentric frequency, root mean square frequency and frequency standard deviation, in order to obtain the frequency domain characteristics of the vibration signal, the time domain signal can be subjected to Fourier transform FFT to obtain a frequency domain signal, and the specific expression is as follows:
mean frequency
Figure QLYQS_14
Center of gravity frequency>
Figure QLYQS_15
Root mean square frequency
Figure QLYQS_16
Frequency standard deviation->
Figure QLYQS_17
Wherein f is k The frequency value corresponding to the kth spectral line is s (k), and the corresponding frequency amplitude value is s (k);
k＝1,2,...,N fft ，N fft is the length of the spectral signal;
frequency characteristic F 15 Reflecting the magnitude of the frequency vibration energy, F 16 ～F 17 Reflecting the change of the position of the main frequency band, F 18 Reflecting the degree of spectral dispersion.
3. The early failure diagnosis method for a rolling bearing according to claim 1, characterized in that: in step 2, smoothing is performed by using a gaussian denoising algorithm, namely, a gaussian weighted moving average filter, and the formula is as follows:
when the length of the sliding window is n, and n is an even number or an odd number:
Figure QLYQS_18
where G is a weight vector satisfying a Gaussian distribution, f i ' is the feature of the ith feature after the smoothing of the gaussian weighted moving average filter, and the expression of the gaussian weight vector G is:
Figure QLYQS_19
/>
where A is the normalized coefficient, a is the width of the Gaussian denoising window,
Figure QLYQS_20
and the formula of the evolution cumulative sum algorithm is as follows:
Figure QLYQS_21
wherein
Figure QLYQS_22
The ith characteristic of the bearing during initial operation is averaged.
4. The early failure diagnosis method for a rolling bearing according to claim 1, characterized in that: the formula of the pearson correlation coefficient in step 3 is as follows:
Figure QLYQS_23
where k is the length of the feature x (i),
Figure QLYQS_24
and &>
Figure QLYQS_25
Are the mean values of x (i) and y (i), respectively; of Pearson's correlation coefficientThe value range is [ -1,1 [)](ii) a The closer the absolute value of the correlation coefficient is to 1, the higher the correlation degree of x and y is; the closer the absolute value of the correlation coefficient is to 0, the lower the correlation of x with y. />
CN202210482268.8A 2022-05-05 2022-05-05 Early fault diagnosis method for rolling bearing Active CN115046764B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210482268.8A CN115046764B (en) 2022-05-05 2022-05-05 Early fault diagnosis method for rolling bearing

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210482268.8A CN115046764B (en) 2022-05-05 2022-05-05 Early fault diagnosis method for rolling bearing

Publications (2)

Publication Number Publication Date
CN115046764A CN115046764A (en) 2022-09-13
CN115046764B true CN115046764B (en) 2023-04-07

Family

ID=83157036

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210482268.8A Active CN115046764B (en) 2022-05-05 2022-05-05 Early fault diagnosis method for rolling bearing

Country Status (1)

Country Link
CN (1) CN115046764B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116124460B (en) * 2022-12-26 2024-01-30 江西理工大学 Bearing life prediction method and system based on health index construction
CN116738372B (en) * 2023-08-15 2023-10-27 昆仑数智科技有限责任公司 Rolling bearing fault diagnosis method, device and equipment for refining centrifugal pump

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101331579B1 (en) * 2013-07-16 2013-11-20 (주) 퓨처파워텍 Automatic control system for diagnosis failure and controlling remaining life by pearson correlation coefficient analysis
CN103954450A (en) * 2014-05-19 2014-07-30 重庆交通大学 Bearing life degradation performance evaluation index construction method based on main component analysis
CN105550700B (en) * 2015-12-08 2019-04-09 国网山东省电力公司电力科学研究院 A kind of time series data cleaning method based on association analysis and principal component analysis
CN106571016B (en) * 2016-11-03 2018-08-07 西安交通大学 A kind of mechanical initial failure method of discrimination based on alarm times saltus step trigger mechanism
CN112633331B (en) * 2020-12-07 2024-07-12 东南大学 Thermal power plant auxiliary machine health state assessment method based on self-encoder
CN113420691A (en) * 2021-06-30 2021-09-21 昆明理工大学 Mixed domain characteristic bearing fault diagnosis method based on Pearson correlation coefficient
CN113901993A (en) * 2021-09-24 2022-01-07 上海海事大学 Fault diagnosis method based on PCCs secondary feature optimization
CN113927371A (en) * 2021-11-05 2022-01-14 太原科技大学 Cutter wear prediction method based on multi-sensor feature fusion

Also Published As

Publication number Publication date
CN115046764A (en) 2022-09-13

Similar Documents

Publication Publication Date Title
CN115046764B (en) Early fault diagnosis method for rolling bearing
Shi et al. Rolling bearing initial fault detection using long short-term memory recurrent network
Wen et al. Graph modeling of singular values for early fault detection and diagnosis of rolling element bearings
CA2969414C (en) Apparatus and method for monitoring a device having a movable part
CN108181105B (en) Rolling bearing fault pre-diagnosis method and system based on logistic regression and J divergence
CN109211548A (en) A kind of mechanical failure diagnostic method
Chen et al. A sparse multivariate time series model-based fault detection method for gearboxes under variable speed condition
Zhang et al. Wavelet energy transmissibility function and its application to wind turbine bearing condition monitoring
CN112629585A (en) Equipment on-line monitoring method and device based on multi-dimensional parameter estimation
Dong et al. A frequency-shifted bispectrum for rolling element bearing diagnosis
CN112182912B (en) Manufacturing equipment spindle bearing health assessment method based on probability description and spectrum analysis
CN111881594A (en) Non-stationary signal state monitoring method and system for nuclear power equipment
CN114841580A (en) Generator fault detection method based on hybrid attention mechanism
CN115220396A (en) Intelligent monitoring method and system for numerical control machine tool
CN112964470A (en) Method for detecting early failure of rolling bearing based on bearing health index
CN113051092B (en) Fault diagnosis method based on optimized kernel density estimation and JS divergence
CN107862175B (en) Factory building vibration multi-scale analysis method
CN116226719A (en) Bearing fault diagnosis method based on multidimensional steady-state vibration characteristics and related components
CN116499742A (en) Rotary machine fault monitoring method based on ensemble learning
CN112857806B (en) Bearing fault detection method based on moving window time domain feature extraction
Zhao et al. Incipient chatter fast and reliable detection method in high-speed milling process based on cumulative strategy
CN113436645A (en) Electromechanical system fault on-line monitoring acoustic processing method under complex noise environment
CN113654651A (en) Strong robust signal early degradation feature extraction and equipment running state monitoring method
CN112287775A (en) Motor fault diagnosis method and device and related components
CN112595515A (en) Power shafting bearing fault detection method and system

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