CN112347653A

CN112347653A - Degradation feature extraction method of bearing and terminal equipment

Info

Publication number: CN112347653A
Application number: CN202011256663.1A
Authority: CN
Inventors: 李洪儒; 李耀龙; 于贺; 许葆华
Original assignee: Army Engineering University of PLA
Current assignee: Army Engineering University of PLA
Priority date: 2020-11-11
Filing date: 2020-11-11
Publication date: 2021-02-09

Abstract

The invention is suitable for the technical field of performance degradation analysis, and provides a method and a device for extracting degradation characteristics of a bearing, wherein the method comprises the following steps: acquiring first characteristic values and second characteristic values of a sample bearing at each preset degradation moment, arranging the first characteristic values in a time sequence to form a first characteristic sequence, and arranging the second characteristic values in the time sequence to form a second characteristic sequence; and determining the fusion characteristics of the sample bearing according to the two time characteristic sequences. In the bearing degradation process, certain two characteristic values of the bearing can show a synergistic relationship, the two characteristics of the sample bearing are organically fused by adopting the method to obtain the fusion characteristic of the sample bearing, the fusion characteristic can show a consistent trend in a certain period, the commonality is obvious, the service life of the bearing is predicted according to the fusion characteristic, and the prediction accuracy is improved.

Description

Degradation feature extraction method of bearing and terminal equipment

Technical Field

The invention belongs to the technical field of performance degradation analysis, and particularly relates to a degradation characteristic extraction method of a bearing and terminal equipment.

Background

The bearing is one of the common parts in mechanical equipment, and whether the bearing is healthy or not directly affects the reliability and the comprehensive benefit of the mechanical equipment, so an effective maintenance scheme needs to be established to ensure the safe and stable operation of the mechanical equipment. The service life of the bearing is effectively predicted, and a reasonable maintenance scheme is formulated to maintain the bearing, so that the safe operation of the bearing can be ensured, and accidents are avoided. As a precondition and a key basis for bearing life prediction, how to effectively extract degradation characteristics of the bearing plays a crucial role in bearing life prediction.

In the prior art, the degradation characteristics of the bearing are generally divided into direct characteristics and indirect characteristics obtained based on a fusion algorithm, but the direct characteristics and the indirect characteristics lack the trend of consistency and have unobvious commonalities, so that the prediction accuracy of the service life of the bearing is not high.

Disclosure of Invention

In view of this, the embodiment of the invention provides a method for extracting degradation features of a bearing and a terminal device, so as to solve the problems that in the prior art, direct features and indirect features are in a trend of lacking consistency, commonalities are not obvious, and the life prediction of the bearing is complex and inaccurate.

A first aspect of an embodiment of the present invention provides a method for extracting degradation characteristics of a bearing, including:

acquiring first characteristic values and second characteristic values of a sample bearing at each preset degradation moment, arranging the first characteristic values in a time sequence to form a first characteristic sequence, and arranging the second characteristic values in the time sequence to form a second characteristic sequence;

determining a third characteristic sequence and a fourth characteristic sequence according to the first characteristic sequence and the second characteristic sequence; the third characteristic sequence and the fourth characteristic sequence have the same single integer order;

and determining the target degradation moment according to the third characteristic sequence and the fourth characteristic sequence, and determining the fusion characteristic of the sample bearing according to the third characteristic sequence and the fourth characteristic sequence.

A second aspect of an embodiment of the present invention provides a degradation feature extraction device for a bearing, including:

the first calculation module is used for acquiring first characteristic values and second characteristic values of the sample bearing at each preset degradation moment, arranging the first characteristic values in time sequence to form a first characteristic sequence, and arranging the second characteristic values in time sequence to form a second characteristic sequence;

the second calculation module is used for determining a third characteristic sequence and a fourth characteristic sequence according to the first characteristic sequence and the second characteristic sequence; the third characteristic sequence and the fourth characteristic sequence have the same single integer order;

and the result output module is used for determining the target degradation moment according to the third characteristic sequence and the fourth characteristic sequence and determining the fusion characteristic of the sample bearing according to the third characteristic sequence and the fourth characteristic sequence.

A third aspect of the embodiments of the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the steps of the degradation feature extraction method for a bearing as provided in the first aspect of the embodiments of the present invention when executing the computer program.

A fourth aspect of the embodiments of the present invention provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the computer program implements the steps of the method for extracting degradation characteristics of a bearing provided in the first aspect of the embodiments of the present invention.

The embodiment of the invention provides a method for extracting degradation characteristics of a bearing, which comprises the following steps: acquiring first characteristic values and second characteristic values of a sample bearing at each preset degradation moment, arranging the first characteristic values in a time sequence to form a first characteristic sequence, and arranging the second characteristic values in the time sequence to form a second characteristic sequence; and determining the fusion characteristics of the sample bearing according to the two time characteristic sequences. In the bearing degradation process, certain two characteristic values of the bearing can show a synergistic relationship, the two characteristics of the sample bearing are organically fused by adopting the method to obtain the fusion characteristic of the sample bearing, the fusion characteristic can show a consistent trend in a certain period, the commonality is obvious, the service life of the bearing is predicted according to the fusion characteristic, and the prediction accuracy is improved.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.

Fig. 1 is a schematic flow chart of an implementation of a method for extracting degradation characteristics of a bearing according to an embodiment of the present invention;

FIG. 2 shows an energy profile E provided by an embodiment of the present invention₁、E₂And E₃A schematic diagram of (a);

FIG. 3 shows an energy profile E provided by an embodiment of the present invention₄、E₅And E₆A schematic diagram of (a);

FIG. 4 shows an energy profile E provided by an embodiment of the present invention₇A schematic diagram of (a);

FIG. 5 is a schematic diagram of complexity of an emulation signal as a function of signal-to-noise ratio provided by an embodiment of the present invention;

FIG. 6 shows the results of the Logistic map model and a LLE thereof according to an embodiment of the present invention;

FIG. 7 is a comparison graph of results of six complexities applied to a Logistic map model according to an embodiment of the present invention;

FIG. 8 provides a bearing life cycle test rig for the US IMS center for an embodiment of the present invention;

FIG. 9 is a PRONOSTIA test stand at the FEMTO-ST research center, France, provided by an embodiment of the present invention;

FIG. 10 is an ABLT-1A type bearing tester of the national detection laboratory of Hangzhou bearing test research center;

FIG. 11 shows RMS values of test set 1 according to an embodiment of the invention²And a plot comparison of sample entropy;

FIG. 12 shows RMS values of test set 2 according to an embodiment of the invention²And a plot comparison of sample entropy;

FIG. 13 shows RMS values of test set 3 according to an embodiment of the invention²And a plot comparison of sample entropy;

FIG. 14 shows RMS values of test set 4 according to an embodiment of the invention²And a plot comparison of sample entropy;

FIG. 15 shows RMS values of test set 5 provided in accordance with an embodiment of the invention²And a plot comparison of sample entropy;

FIG. 16 is a graph of the fusion characteristics of test set 1 provided by an embodiment of the present invention;

FIG. 17 is a graph of a fusion signature of test set 2 provided by an embodiment of the present invention;

FIG. 18 is a graph of the fusion characteristics of test set 3 provided by an embodiment of the present invention;

FIG. 19 is a graph of the fusion characteristics of test set 4 provided by an embodiment of the present invention;

FIG. 20 is a graph of the fusion characteristics of test set 5 provided by an embodiment of the present invention;

FIG. 21 shows a detail 2-1 data fusion feature and RMS provided in accordance with an embodiment of the present invention²A graph comparing the curves of (1);

FIG. 22 RMS of Bearing2-1 data provided by an embodiment of the invention²And a curve diagram of fusion characteristics is obtained after sample entropy is subjected to PCA fusion;

FIG. 23 RMS of Bearing2-1 data provided by an embodiment of the invention²And a curve diagram of fusion characteristics is obtained after the sample entropy is fused by Isomap;

fig. 24 is a schematic view of a degradation feature extraction device of a bearing according to an embodiment of the present invention;

fig. 25 is a schematic diagram of a terminal device according to an embodiment of the present invention.

Detailed Description

In the following description, for purposes of explanation and not limitation, specific details are set forth, such as particular system structures, techniques, etc. in order to provide a thorough understanding of the embodiments of the invention. It will be apparent, however, to one skilled in the art that the present invention may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present invention with unnecessary detail.

In order to explain the technical means of the present invention, the following description will be given by way of specific examples.

Referring to fig. 1, an embodiment of the present invention provides a method for extracting degradation characteristics of a bearing, including:

step S101: the method comprises the steps of obtaining first characteristic values and second characteristic values of a sample bearing at each preset degradation moment, arranging the first characteristic values in a time sequence to form a first characteristic sequence, and arranging the second characteristic values in the time sequence to form a second characteristic sequence.

The bearing is one of the most extensive and most volatile parts in the rotating machinery, the operation state of the bearing is directly related to the overall performance of equipment, and the prediction or fault prediction of the residual life of the bearing is of great significance to the continuous and stable operation of the mechanical equipment. The bearing characteristic shows a certain change trend along with the use time, first characteristic values and second characteristic values of each preset degradation moment of the sample bearing are obtained, the first characteristic values are arranged according to the time sequence to form a first characteristic sequence, and the second characteristic values are arranged according to the time sequence to form a second characteristic sequence.

In some embodiments, the first characteristic value is a value of an energy characteristic and the second characteristic value is a value of a complexity characteristic.

The existing characteristic classification comprises two types of traditional degradation characteristics, one is an energy characteristic with an ascending trend, which reflects the energy change of a bearing in the operation process, and the RMS is used²Is a typical (RMS) Root Mean Square, which is widely used in degradation state identification and residual life prediction of bearings with its good performance. Another class is complexity features with a decreasing trend, such as approximate entropy, sample entropy, permutation entropy, etc. Characterizing energy versus complexityAnd the degradation condition of the bearing can be comprehensively reflected by fusion. Meanwhile, the two have opposite change trends, and the consistent change trend can be presented after fusion.

In some embodiments, the energy characteristic may include RMS²The complexity feature may include sample entropy.

1. Energy characteristics

The energy characteristics reflect the energy change of the bearing in the operation process, the square of the amplitude characteristics of the signals can reflect the energy, common amplitude characteristics can comprise square root amplitude, root mean square value (RMS), absolute mean value, peak-to-peak value, maximum value, minimum value and spectrum average value, and the calculation formulas are respectively as follows:

E₄＝max(x_i)-min(x_i)

E₅＝max{|x_i|}

E₆＝-min{x_i}

wherein E is₁Is the square root amplitude, E₂Is the root mean square value, E₃Is an absolute mean value, E₄Is the peak-to-peak value, E₅Is a maximum value, E₆Is a minimum value and E₇Is the average value of the frequency spectrum; x is the number of_iIs the ith measurement in the signal, s (k) is the ith amplitude in the signal domain; i is 1,2, …, N.

From FIG. 2 toIn FIG. 4, E₁、E₂、E₃、E₄、E₅、E₆And E₇The trend is similar. E can be obtained by ADF inspection (Augmented Dickey-Fuller, Unit root inspection)₁、E₂、E₃And E₇Is 2-order simple, E₄、E₅、E₆Is 1 order single integer. By definition, the same order monographs can be tested for E-G (E-G) synergy, and the results are shown in Table 1. At significance level of 0.1, E₁、E₂、E₃Have a synergistic relationship between them. E₄、E₅、E₆Have a synergistic relationship between them. That is, E₁～E₃Having the same tendency to change, E₄～E₆With the same trend of variation. E₇Although representing some energy characteristic, there is no co-integration with RMS at the current significance level because E₇The Fourier transform exists in the process of obtaining, and the Fourier transform has the problems of aliasing, leakage and the like. E₄～E₆Characterize the extrema of each set of signals, with poor stability, and E₁～E₃The average energy of each group of signals is represented, and the stability is strong. Since the RMS is widely used in industry, the amplitude characteristic may be represented by the RMS, and thus, the energy characteristic may include the RMS²。

TABLE 1E-G synergy test results for energy characteristics

2. Complexity characterization

The complexity characteristics reflect the complexity of the signal, and the common complexity characteristics include approximate entropy, sample entropy, fuzzy entropy, shannon entropy, permutation entropy, L-Z complexity and the like. In order to test the performance of each complexity feature, the same parameters are set to be consistent in the calculation process so as to reduce the influence of the parameters on the result. Table 2 lists the selection of complexity parameters.

Sample entropy is an improved method of approximating entropy, typically, the embedding dimension m takes 1 or 2; the similarity margin r is usually chosen to be 0.2std, where std is the standard deviation of the original signal. The shannon entropy is calculated by dividing data, and setting an extremum to divide the data into 50 intervals. The permutation entropy is similar to the shannon entropy in the calculation process, the embedding dimension of the permutation entropy is different from the sample entropy and the approximate entropy, the larger the embedding dimension of the permutation entropy is, the more accurate the permutation entropy is, but the longer the time consumption is, the calculation amount of the permutation entropy is the multiplication of the embedding dimension, and the embedding dimension of the permutation entropy is set to be 6 by considering.

TABLE 2 parameter selection for complexity

In order to test the performance of complexity, a simulation signal is set, and the formula of the simulation signal can be s (t) ═ x (t) + e (t); where x (t) is a sine signal (2 pi × 10t), and e (t) is white gaussian noise. The sampling frequency was 10000Hz, and the sampling time was 1 s. By varying the intensity of the noise and thus the signal-to-noise ratio, the variation in the complexity of the simulated signal is shown in fig. 5. In theory, complexity should increase with increasing noise. It can be seen that the shannon entropy and the permutation entropy are not completely monotonous, which indicates that the performance of the shannon entropy and the permutation entropy is not good enough.

To further test performance, the complexity is tested using more general signals. The Logistic map model (Logistic map) is a typical non-linear system, the expression of which can be simply expressed as x_n+1＝μx_n(1-x_n) Wherein x is_nIs data of n points, and μ is a coefficient. The model comprises a large number of periods and chaotic signals, the complexity of the periodic signals is 0, and the complexity of the chaotic signals is high. FIG. 6 shows 2.5<μ<The 4 th Logistic map model result and the corresponding maximum Lyapunov index (maximum Lyapunov index, LLE, the largest Lyapunov exponent). The LLE may reflect the complexity of the signal being produced, LLE<When 0, the signal is a periodic signal, when LLE is 0, the signal is a bifurcation point, LLE>And when 0, the signal is a chaotic signal. The 6 complexity features were introduced into the Logistic map model, the results of which are shown in FIG. 7Shown in the figure. As can be seen from fig. 7, shannon entropy and permutation entropy have errors in the measurement of the periodic signal. The fuzzy entropy is biased when the measure mu is 3.5, which is a problem caused by the fuzzy membership of the fuzzy entropy. The L-Z complexity is biased around 3.6 due to the coarse-grained process of L-Z complexity in the calculation process.

From the above analysis, the approximate entropy and the sample entropy perform well in six complexities. Meanwhile, as an improved algorithm of approximate entropy, the sample entropy does not contain comparison of self data segments during calculation, the advantage of the method is that the time sequence length is less depended on, and the consistency of the result is better. Thus, the sample entropy performs best among these six complexities, and the complexity features may include the sample entropy.

Step S102: determining a third characteristic sequence and a fourth characteristic sequence according to the first characteristic sequence and the second characteristic sequence; wherein the third signature sequence and the fourth signature sequence have the same single integer order.

In some embodiments, step S102 may include:

determining the single integer order of the first characteristic sequence and the single integer order of the second characteristic sequence;

if the single integer order of the first feature sequence is the same as the single integer order of the second feature sequence, the first feature sequence is the third feature sequence, and the second feature sequence is the fourth feature sequence;

if the single integer order of the first characteristic sequence is different from the single integer order of the second characteristic sequence, removing the last first characteristic value in the first characteristic sequence to obtain a first intermediate sequence, taking the first intermediate sequence as a new first characteristic sequence, removing the last second characteristic value in the second characteristic sequence to obtain a second intermediate sequence, taking the second intermediate sequence as a new second characteristic sequence, and continuously executing the step of determining the single integer order of the first characteristic sequence and the single integer order of the second characteristic sequence.

In some embodiments, a single integer order of the first signature sequence and a single integer order of the second signature sequence may be determined by a unit root test.

Step S103: and determining a target degradation moment according to the third characteristic sequence and the fourth characteristic sequence, and determining the fusion characteristic of the sample bearing according to the third characteristic sequence and the fourth characteristic sequence.

In some embodiments, the third feature sequence includes a plurality of first feature values, the fourth feature sequence includes a plurality of second feature values, and step S103 may include:

determining whether the third signature sequence and the fourth signature sequence have a co-integration relationship;

if the third feature sequence and the fourth feature sequence have a co-integration relationship, determining a co-integration vector between the third feature sequence and the fourth feature sequence according to the third feature sequence and the fourth feature sequence, and determining a fusion feature of the sample bearing according to the co-integration vector; the value of the preset degradation moment corresponding to the last first characteristic value in the third characteristic sequence is a target degradation moment;

if the third feature sequence and the fourth feature sequence do not have a co-integration relationship, removing a last first feature value in the third feature sequence to obtain a third intermediate sequence, taking the third intermediate sequence as a new third feature sequence, removing a last second feature value in the fourth feature sequence to obtain a fourth intermediate sequence, taking the fourth intermediate sequence as a new fourth feature sequence, and continuing to execute the step of determining whether the third feature sequence and the fourth feature sequence have the co-integration relationship.

In some embodiments, an E-G co-integration test may be used to determine whether the third signature sequence and the fourth signature sequence have a co-integration relationship.

In some embodiments, an E-G co-integration test may be used to determine a co-integration vector between the third signature sequence and the fourth signature sequence.

Vectors consisting of n sets of d-order monomorphic sequences according to the definition of co-integrationy_t＝[y_1t,y_2t,…,y_nt]If there is a vector β ═ β₁,β₂,…,β_n]So that the linear combination β y_t＝[β₁y_1t+β₂y_2t+…+β_ny_nt]Is d-b order simple, wherein b>0, then y is considered_t＝[y_1t,y_2t,…,y_nt]Is a (d, b) order co-integration, denoted as y_tCI (d, b), the vector β is called a co-integration vector. A common co-integration relationship is CI (1,1), i.e. n sets of 1 st order single integer sequences have a co-integration, which combines the residuals ε linearly according to a co-integration vector_tIs a stationary sequence.

For example, in the embodiment of the present invention, the feature vector y composed of the third feature sequence and the fourth feature sequence 2 groups of single sequences_t＝[y_1t,y_2t]There is one feature vector β ═ β₁,β₂]So that the linear combination β y_t＝[β₁y_1t,β₂y_2t]And if the order is d-b simple integer, the third characteristic sequence and the fourth characteristic sequence have a synergistic relation. The characteristic vector beta ═ beta can be obtained by an E-G co-integration test method₁,β₂]With a fusion characteristic of beta₁y_1t+β₂y_2t。

The method for extracting the degradation features of the bearing, provided by the embodiment of the invention, is used for acquiring the first feature value and the second feature value of the sample bearing at each preset degradation moment, respectively forming the first feature sequence and the second feature sequence, further determining the fusion features of the sample bearing according to the first feature sequence and the second feature sequence, organically fusing the first feature and the second feature together by the fusion features, presenting a consistent trend within a certain period, and having obvious commonality.

In some embodiments, determining the fused features of the sample bearing from the co-integration vector comprises:

and determining a co-integration vector between the third characteristic sequence and the fourth characteristic sequence according to the third characteristic sequence and the fourth characteristic sequence, and carrying out linear combination on the third characteristic sequence and the fourth characteristic sequence according to the co-integration vector to obtain the fusion characteristic of the sample bearing.

The above method is further described with reference to specific examples below:

and selecting multiple groups of bearing full-life data to verify the method, wherein the data is divided into public data and self-service data.

Common public data include a life test data set and an IEEE PHM 2012 cognitive change data set of an Intelligent Maintenance Systems (IMS) center (abbreviated as the IMS center). Referring to fig. 8, in the united states IMS center, four bearings are mounted on a shaft during testing, the bearing model is Rexnord-ZA2115 angular contact bearing, the rotating speed of the shaft is fixed to 2000r/min, the radial load is 26690N, and two 353B33 type acceleration sensors are mounted on each bearing. The sampling frequency of the test bed is 20kHz, each group of sampling samples comprises 20480 data points, and the sampling interval is 10 min.

In the embodiment of the invention, Bearing1-4 data in a full-life test data set of the American IMS center is selected as a test set 1, and the failure mode is rolling element failure; bearing1-3 was used as test set 2, and its failure mode was inner ring failure. Test set 1 and test set 2 each contained 2500 sets of data.

The IEEE PHM 2012 protective Challenge data set was acquired on a PRONOSTIA test stand at the research center of FEMTO-ST, france, with reference to fig. 9, with a sampling frequency of 25.6kHz, a sampling time of 0.1s, and a sampling interval of 10 s. In the IEEE PHM 2012 scientific Challenge data set, Bearing1-1PHM data is selected as a test set 3, and Bearing1-2PHM data is selected as a test set 4. Test set 3 and test set 4 contain 2803 and 871 sets of data, respectively.

While the common data is adopted, the embodiment of the invention also adopts bearing full-life acceleration test data of a national detection laboratory of Hangzhou bearing test research center. The test was carried out on an ABLT-1 type A bearing tester, as shown in FIG. 10. A6204 type bearing is adopted in the test, the sampling interval is 10min, the sampling frequency is 25.6kHz, and the sampling time is 1 s. In the embodiment of the invention, a group of data in the bearing full-life acceleration test data of a national detection laboratory of Hangzhou bearing test research center is selected as a test set 5, and the data set comprises 960 groups of data.

First, RMS and sample entropy of each test set are extracted, and referring to fig. 11 to 15, degradation curves of each test set are differently represented. Test set 1 showed a "healing" phenomenon with RMS increasing first and then decreasing. The RMS of test set 2 showed a long-term plateau and finally a rapid rise, at which time a "healing" phenomenon was also observed. From fig. 12, a step appears around 160 RMS, and the original signal is analyzed, where there is a shutdown, which shows that the startup and shutdown has an influence on the working condition, and the startup and shutdown frequency should be reduced when a full-life test is performed. The RMS of test set 3 rose first smoothly and then rapidly, with sample entropy rising and falling again around 2500 groups. The RMS of the test set 4 appeared stable in the early and middle stages, but the short-term fluctuation was strong, and the RMS rose rapidly in the failure stage. The RMS of test set 5 appeared to plateau between early and mid-term periods and then rose slowly to eventually fail.

From the above, it can be obtained that the RMS and the sample entropy of different test sets have characteristics, do not have a uniform change rule, and lack consistency.

RMS testing of each test set by E-G²By means of the coordination with the sample entropy, the RMS of each test set can be obtained²And the partial sequences of the sample entropy are consistent. The degradation features of each test set are extracted according to the degradation feature extraction method of the bearing in the above-described embodiment of the invention, as shown in fig. 16 to 20. As can be seen from the figure, the degradation characteristics of each test set have obvious two-stage property. It shows stationarity in the early and middle stages of the whole life. This feature appears non-stationary when the bearing is in the failure period and has some monotonicity. The RMS and sample entropy of each test set represent different bearing degradation processes, but after the degradation feature extraction method of the bearing provided by the embodiment of the invention is fused based on the co-arrangement theory, the fused features represent consistencyTherefore, the method for extracting the degradation characteristics of the bearing provided by the embodiment of the invention can unify the evolution process of different bearings, has generality, and simultaneously reduces the long-term trend fluctuation of RMS and sample entropy in the early and middle stages of the full life.

In the full-life test data set of the IMS center in the United states, Bearing2-1 data is the most commonly used, and the failure mode is outer ring failure. 982 sets of data of Bearing2-1 were selected for analysis, and RMS values of Bearing2-1 data are shown in FIG. 21²And fusion features, RMS by Bearing2-1 data²From the start of the run to group 510, the RMS curve²The stability is kept, and the bearing is known to be in a normal state; from 510 groups to 700 groups, RMS²Continuously rising, and enabling the bearing to be in a slight fault stage; 700 groups time, RMS²The sudden increase, which may be caused by the protrusions on the friction face, then runs to 823 groups, RMS²Undergoes a phase of descent and re-ascent, mainly due to the reduction of the surface of the local protrusions by the action of continuous friction, and then RMS²The process of descending and ascending is carried out, and the bearing is in a medium fault stage; from 900 to 982 groups, RMS²The rise continues with the bearing in a severe failure stage until failure. This process of descending and re-ascending is known as the "healing" phenomenon, which is common in bearings, and occurs because of the reduction of the surface of the local protrusions by the continuous friction. Then, as the fault deepens, the friction pair surface forms a new bulge again, and the process is repeated. The "healing" phenomenon indicates that the bearing has entered a moderate failure condition, the presence of which, while beneficial for degradation condition identification, also results in RMS²The monotonicity of (a) is reduced, which is not beneficial to the prediction of the bearing.

The fusion characteristic curve of the Bearing2-1 data shows that the data has two sections. From the beginning, running to around 914 groups, it has stationarity; from group 914 to the last failure, it rises rapidly. RMS when a bearing is in a normal condition²And the entropy of the sample is stable. When the bearing is in a light failure phase, protrusions appear on the friction surface, resulting in an increase in energy, RMS²The value rises; at the same time, the bumps may cause signalsThe periodicity of (2) increases and the complexity decreases. When the bearing is in a "healing" condition, continued friction smoothes the protrusions, which results in a reduction in energy, RMS²Descending; as the bumps become smooth, the periodicity decreases accordingly, resulting in increased complexity. In summary, prior to group 914, RMS²Keeping synchronization with the sample entropy, thus making the linear combination of the two sequences a stationary sequence. When the bearing is close to failure, the energy is increased sharply, and the protrusions on the surface of the friction pair are increased sharply, and the effect is not obvious although the effect is smooth. Because each bump can form a periodic signal, the integral vibration signal is the signal superposition caused by each bump, the signal periodicity is not obvious at the moment, and the complexity characteristic is not obviously reduced. At this point, the synergy of the two sequences is lost and the fusion characteristics rise significantly at this stage. The target degradation moment can be regarded as a turning point of the change from the local fault to the multi-point fault, and when the target degradation moment is reached, the bearing is not far away from the failure, and the residual life prediction of the bearing should be carried out.

And evaluating the prediction performance of the fusion characteristics by monotonicity, robustness and trend. Monotonicity measures the degree of monotonous fusion characteristics, and smoothing processing is required to be carried out before solving, so that the influence of noise is reduced. Robustness measures the degree of stability of the fused features, thereby reducing uncertainty of the prediction result. Trending refers to the degree to which features correlate with runtime. The value ranges of the three indexes are [0,1], and the closer to 1, the better the corresponding performance of the characteristics is. The Bearing2-1 data and the predicted performance indicators for the 5 test sets were determined separately, as shown in table 3 below.

The fusion features should be normalized during the calculation, and the influence of the stationary phase of each feature is ignored during the calculation of monotonicity and trend. For example, the smoothing method employs a gaussian filtering method, window length 20.

As can be seen from Table 3, the degeneration characteristics based on the co-integration theory have obvious non-subtractive characteristics. Relative to RMS²And sample entropy, which is better monotonicity in non-stationary phases. Fusing features with RMS in robustness and trending²And entropy of the sampleAt the same level. In fact, the method for extracting the degradation characteristic of the bearing based on the co-integration theory in the embodiment of the invention is a fusion algorithm. Typical fusion algorithms include linear and nonlinear dimensionality reduction methods, of which Principal Component Analysis (PCA) and hierarchical Feature Mapping (Isomap) are typical. Taking Bearing2-1 data as an example, RMS is measured²The result shown in FIG. 22 is obtained by PCA fusion of the sample entropy and RMS²And the result is obtained by Isomap fusion with sample entropy as shown in figure 23. As can be seen from fig. 22 and 23, the features after PCA fusion and Isomap fusion are very similar to the sample entropy. RMS overall²In contrast to the trend of sample entropy, features fused with Isomap via PCA retain their common trend. Unlike PCA and Isomap, the embodiment of the present invention provides a fusion feature that extracts the places where the trends are different from each other, and eliminates the common trend, for example, the front 914 data of Bearing2-1 data is subjected to the integration and fusion to eliminate the common trend. However, the fused features also ignore the trend information common to the two original features, so that the original features should be considered on the basis of the fused features to realize accurate prediction of the service life of the bearing when the service life of the bearing is predicted.

TABLE 3 predicted Performance indices for fusion features of respective test sets

From the above, the fusion characteristics of the bearing determined by the method have good two-stage performance, and can reflect different characteristics of the bearing in the early stage, the middle stage and the failure stage. Meanwhile, the fusion feature can reduce the long-term fluctuation of RMS and sample entropy before and during degeneration. Secondly, the fusion characteristic has generality, and can unify the life cycle data of different rolling bearings to obtain a consistent evolution process. And thirdly, compared with the RMS and the sample entropy, the monotonicity of the fusion characteristics in a non-stationary stage is good, and the fault prediction capability is better.

It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

Referring to fig. 24, an embodiment of the present invention further provides a degradation feature extraction device for a bearing, including:

the first calculating module 201 is configured to obtain first characteristic values and second characteristic values of the sample bearing at each preset degradation time, arrange the first characteristic values in a time sequence to form a first characteristic sequence, and arrange the second characteristic values in a time sequence to form a second characteristic sequence;

a second calculating module 202, configured to determine a third feature sequence and a fourth feature sequence according to the first feature sequence and the second feature sequence; wherein the third signature sequence and the fourth signature sequence have the same single integer order;

and a result output module 203, configured to determine a target degradation time according to the third feature sequence and the fourth feature sequence, and determine a fusion feature of the sample bearing according to the third feature sequence and the fourth feature sequence.

In some embodiments, the second computing module 202 may include:

a single integer order determining unit, configured to determine a single integer order of the first feature sequence and a single integer order of the second feature sequence;

a first determining unit, configured to determine that the first feature sequence is the third feature sequence and the second feature sequence is the fourth feature sequence if the single integer number of the first feature sequence is the same as the single integer number of the second feature sequence;

a second determining unit, configured to remove a last first feature value in the first feature sequence to obtain a first intermediate sequence if the single integer number of the first feature sequence is different from the single integer number of the second feature sequence, take the first intermediate sequence as a new first feature sequence, remove a last second feature value in the second feature sequence to obtain a second intermediate sequence, take the second intermediate sequence as a new second feature sequence, and continue to perform the step of determining the single integer number of the first feature sequence and the single integer number of the second feature sequence.

In some embodiments, the third feature sequence includes a plurality of first feature values, the fourth feature sequence includes a plurality of second feature values, and the result output module 203 may include:

a co-integration relation determining unit, configured to determine whether the third feature sequence and the fourth feature sequence have a co-integration relation;

a third judging unit, configured to determine, if the third feature sequence and the fourth feature sequence have a co-integration relationship, a co-integration vector between the third feature sequence and the fourth feature sequence according to the third feature sequence and the fourth feature sequence, and determine a fusion feature of the sample bearing according to the co-integration vector; the value of the preset degradation moment corresponding to the last first characteristic value in the third characteristic sequence is a target degradation moment;

a fourth determining unit, configured to, if the third feature sequence and the fourth feature sequence do not have a co-integration relationship, remove a last first feature value in the third feature sequence to obtain a third intermediate sequence, take the third intermediate sequence as a new third feature sequence, remove a last second feature value in the fourth feature sequence to obtain a fourth intermediate sequence, take the fourth intermediate sequence as a new fourth feature sequence, and continue to perform the step of determining whether the third feature sequence and the fourth feature sequence have a co-integration relationship.

It is obvious to those skilled in the art that, for convenience and simplicity of description, the foregoing division of the functional units and modules is merely used as an example, and in practical applications, the foregoing function distribution may be performed by different functional units and modules as needed, that is, the internal structure of the terminal device is divided into different functional units or modules to perform all or part of the above-described functions. Each functional unit and module in the embodiments may be integrated in one processing unit, or each unit may exist alone physically, or two or more units are integrated in one unit, and the integrated unit may be implemented in a form of hardware, or in a form of software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working processes of the units and modules in the above-mentioned apparatus may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.

Fig. 25 is a schematic block diagram of a terminal device according to an embodiment of the present invention. As shown in fig. 25, the terminal device 4 of this embodiment includes: one or more processors 40, a memory 41, and a computer program 42 stored in the memory 41 and executable on the processors 40. The processor 40, when executing the computer program 42, implements the steps in the above-described degradation feature extraction embodiments of the respective bearings, such as the steps S101 to S103 shown in fig. 1. Alternatively, the processor 40, when executing the computer program 42, implements the functions of the modules/units in the above-described degraded characteristic extraction apparatus embodiment of the bearing, for example, the functions of the modules 201 to 203 shown in fig. 24.

Illustratively, the computer program 42 may be partitioned into one or more modules/units that are stored in the memory 41 and executed by the processor 40 to accomplish the present application. The one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, which are used to describe the execution process of the computer program 42 in the terminal device 4. For example, the computer program 42 may be divided into a first computing module, a second computing module, and a result output module.

the second calculation module is used for determining a third characteristic sequence and a fourth characteristic sequence according to the first characteristic sequence and the second characteristic sequence; wherein the third signature sequence and the fourth signature sequence have the same single integer order;

and the result output module is used for determining a target degradation moment according to the third characteristic sequence and the fourth characteristic sequence and determining the fusion characteristic of the sample bearing according to the third characteristic sequence and the fourth characteristic sequence.

Other modules or units can be referred to the description of the embodiment shown in fig. 24, and are not described again here.

The terminal device 4 includes, but is not limited to, a processor 40 and a memory 41. Those skilled in the art will appreciate that fig. 25 is only one example of a terminal device and does not constitute a limitation to terminal device 4 and may include more or less components than those shown, or combine certain components, or different components, for example, terminal device 4 may also include an input device, an output device, a network access device, a bus, etc.

The Processor 40 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 41 may be an internal storage unit of the terminal device, such as a hard disk or a memory of the terminal device. The memory 41 may also be an external storage device of the terminal device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, which are provided on the terminal device. Further, the memory 41 may also include both an internal storage unit of the terminal device and an external storage device. The memory 41 is used for storing the computer program 42 and other programs and data required by the terminal device. The memory 41 may also be used to temporarily store data that has been output or is to be output.

In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to the related descriptions of other embodiments for parts that are not described or illustrated in a certain embodiment.

Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

In the embodiments provided in the present application, it should be understood that the disclosed terminal device and method may be implemented in other ways. For example, the above-described terminal device embodiments are merely illustrative, and for example, the division of the modules or units is only one logical function division, and there may be other divisions when actually implemented, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The integrated modules/units, if implemented in the form of software functional units and sold or used as separate products, may be stored in a computer readable storage medium. Based on such understanding, all or part of the flow in the method of the embodiments described above can be realized by a computer program, which can be stored in a computer-readable storage medium and can realize the steps of the embodiments of the methods described above when the computer program is executed by a processor. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like. It should be noted that the computer readable medium may contain other components which may be suitably increased or decreased as required by legislation and patent practice in jurisdictions, for example, in some jurisdictions, computer readable media which may not include electrical carrier signals and telecommunications signals in accordance with legislation and patent practice.

The above-mentioned embodiments are only used for illustrating the technical solutions of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present application and are intended to be included within the scope of the present application.

Claims

1. A method for extracting degradation characteristics of a bearing, comprising:

determining a third characteristic sequence and a fourth characteristic sequence according to the first characteristic sequence and the second characteristic sequence; wherein the third signature sequence and the fourth signature sequence have the same single integer order;

and determining a target degradation moment according to the third characteristic sequence and the fourth characteristic sequence, and determining the fusion characteristic of the sample bearing according to the third characteristic sequence and the fourth characteristic sequence.

2. The method of extracting degradation characteristics of a bearing according to claim 1, wherein the first characteristic value is a value of an energy characteristic, and the second characteristic value is a value of a complexity characteristic.

3. The method of extracting degradation characteristics of a bearing of claim 2, wherein the energy characteristics include RMS²The complexity feature comprises sample entropy.

4. The method of extracting degradation characteristics of a bearing according to claim 1, wherein the determining the third and fourth feature sequences from the first and second feature sequences comprises:

5. The method of extracting degradation characteristics of a bearing according to claim 4, wherein the determining the single integer order of the first signature sequence and the single integer order of the second signature sequence comprises:

and determining the single integer order of the first characteristic sequence and the single integer order of the second characteristic sequence by adopting a unit root checking method.

6. The method of extracting degradation characteristics of a bearing according to claim 1, wherein the third feature sequence includes a plurality of first feature values, and the fourth feature sequence includes a plurality of second feature values;

the determining a target degradation time according to the third feature sequence and the fourth feature sequence, and determining a fusion feature of the sample bearing according to the third feature sequence and the fourth feature sequence includes:

7. The method of extracting degradation features of a bearing of claim 6, wherein the determining whether the third feature sequence and the fourth feature sequence have a co-integration relationship comprises:

and determining whether the third characteristic sequence and the fourth characteristic sequence have a synergistic relationship by adopting an E-G synergistic integration test method.

8. A degradation feature extraction device of a bearing, characterized by comprising:

9. A terminal device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the method for extracting degradation features of a bearing according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the steps of the degradation feature extraction method of a bearing according to any one of claims 1 to 7.