CN110207987B - A method for determining the degradation node of rolling bearing performance degradation - Google Patents

Abstract

The invention discloses a method for judging the degradation node of rolling bearing performance degradation, which relates to the technical field of bearing performance detection. value, generate the maximum power spectral density curve, generate the first derivative curve of the maximum power spectral density, define the decay period node, and determine the predictable interval. In the whole bearing life cycle, the present invention can make the change trend of the extracted performance degradation characteristic quantity of different bearings consistent, and can determine the working stage of the bearing and the node entering the recession stage according to the trend curve, and then determine the predictable interval. In addition, the method can effectively improve the accuracy of the prediction of the remaining life of the rolling bearing, which provides a method basis for ensuring the safety, availability and efficient operation of the system, reducing maintenance costs, and realizing condition repair.

Description

A method for determining the degradation node of rolling bearing performance degradation

technical field

The invention relates to the technical field of bearing performance detection, in particular to a method for determining a node of performance degradation and decline of a rolling bearing.

Background technique

Different bearings have different life values due to their different working conditions and damage types. It is difficult to find an index that reflects the consistency of the change trend of different bearing performance degradation characteristics, which brings difficulties to improve the prediction accuracy of bearing remaining life. The entire working cycle of the bearing can be divided into a running-in stage, a normal stage and a recession stage. The remaining life prediction of the bearing should select the performance decline period as the prediction interval. If the index reflecting the consistency of the change trend of the extracted performance degradation feature quantity cannot be found, there will be no uniform standard for the division of the working cycle of the bearing. Therefore, it is crucial to find an index that reflects the consistency of the change trend of the extracted feature quantities, which can improve the accuracy of bearing residual life prediction and provide a unified standard for the determination of the recession period.

SUMMARY OF THE INVENTION

In order to solve the above-mentioned problems in the background technology, the present invention proposes a method for determining the degradation node of rolling bearing performance degradation. This method can make the change trend of the extracted performance degradation characteristic quantities of different bearings consistent throughout the life cycle. According to the trend curve, the working stage of the bearing and the node entering the recession period can be determined, and then the predictable interval can be determined.

The technical scheme of the present invention is:

A method for judging the degradation node of rolling bearing performance degradation, the specific steps of the method are:

Step 1: Obtain the vibration acceleration data of different monitoring periods, and perform Hilbert transform on the data samples Dataij in the bearing life data set bi to obtain the power spectral density PSD_Dataij of Dataij;

Step 2: Find the maximum value PSD_MAX_Dataij of the power spectral density curve PSD_Dataij of the data Dataij, and form the maximum power spectral density set PMDi={PSD_MAX_Dataij|i=1,2,…,n,j=1,2,…,Ni} , n is the number of bearings included in the bearing life data set bi, and Ni is the number of data samples in bi;

Step 3: use quartic polynomial fitting to perform curve fitting on the maximum power spectral density set PMDi to obtain the fitting curve PPMDi;

Step 4: Differentiate the fitting curve PPMDi to obtain the derivation curve DPPMDi of the fitting curve, which is the change trend of the characteristic quantity;

Step 5: Find the inflection point of the DPPMDi curve, and define the first inflection point as the node where the bearing enters the normal working state, and the second inflection point as the node where the bearing enters the performance decline stage, and determine the predictable interval.

Preferably, the derivation curve DPPMDi is the first derivative curve of the maximum power spectral density.

The beneficial effects of the method for judging the degradation node of rolling bearing performance proposed by the present invention: in the whole bearing life cycle, the change trend of the extracted performance degradation characteristic quantities of different bearings can be made consistent, and the bearing can be determined according to the trend curve. The working stage and the node entering the recession period, and then determine the predictable interval. In addition, the method can effectively improve the accuracy of the prediction of the remaining life of the rolling bearing, which provides a method basis for ensuring the safety, availability and efficient operation of the system, reducing maintenance costs, and realizing condition repair.

Description of drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained according to these drawings without creative efforts.

Fig. 1 is a specific flow chart of a method for judging a node of performance degradation of a rolling bearing provided by the present invention;

FIG. 2 is a characteristic curve diagram of the performance degradation of the bearing throughout the life of the present invention; wherein,

a(1) is the maximum power spectral density curve of bearing1_1 and its fitting curve, a(2) is the first derivative of the fitting curve;

b(1) is the maximum power spectral density curve of bearing1_2 and its fitting curve, and b(2) is the first derivative of the fitting curve;

c(1) is the maximum power spectral density curve of bearing1_5 and its fitting curve, and c(2) is the first derivative of the fitting curve;

d(1) is the maximum power spectral density curve of bearing1_7 and its fitting curve, d(2) is the first derivative of the fitting curve;

e(1) bearing2_1 power spectral density maximum curve and its fitting curve, e(2) is the first derivative of the fitting curve;

f(1) is the maximum power spectral density curve of bearing2_2 and its fitting curve, and f(2) is the first derivative of the fitting curve;

g(1) is the maximum power spectral density curve of bearing2_5 and its fitting curve, and g(2) is the first derivative of the fitting curve;

h(1) is the maximum power spectral density curve of bearing2_6 and its fitting curve, and h(2) is the first derivative of the fitting curve;

i(1) is the maximum power spectral density curve of bearing2_7 and its fitting curve, and i(2) is the first derivative of the fitting curve;

j(1) is the maximum power spectral density curve of bearing3_1 and its fitting curve, and j(2) is the first derivative of the fitting curve;

k(1) is the maximum power spectral density curve of bearing3_2 and its fitting curve, and k(2) is the first derivative of the fitting curve.

Fig. 3 is the characteristic curve diagram of bearing full life under different feature extraction methods of the present invention; wherein,

(a) 1 is the time domain curve diagram of bearing1_1 full life, (a) 2 is the full life time domain curve of bearing1_2;

(b)1 is the bearing1_1HOMMSE curve, (b)2 is the bearing1_2HOMMSE curve;

(c)1 is the bearing1_1RMS curve, (c)2 is the bearing1_2RMS curve;

(d)1 is the curve of bearing1_1EE, (d)2 is the curve of bearing1_2EE;

(e)1 is the bearing1_1PMD graph, (e)2 is the bearing1_2PMD graph.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

1 Experimental method

Referring to FIG. 1 , an embodiment of the present invention provides a method for determining a node of performance degradation of a rolling bearing. The specific steps of the method are as follows:

Step 1: Obtain the vibration acceleration data of different monitoring periods, and perform Hilbert transform on the data samples Dataij in the bearing life data set bi to obtain the power spectral density PSD_Dataij of Dataij;

Step 2: Find the maximum value PSD_MAX_Dataij of the power spectral density curve PSD_Dataij of the data Dataij, and form the maximum power spectral density set PMDi={PSD_MAX_Dataij|i=1,2,…,n,j=1,2,…,Ni} , n is the number of bearings included in the bearing life data set bi, and Ni is the number of data samples in bi;

Step 3: use quartic polynomial fitting to perform curve fitting on the maximum power spectral density set PMDi to obtain the fitting curve PPMDi;

Step 4: derivation of the fitting curve PPMDi to obtain a derivation curve DPPMDi of the fitting curve, which is the variation trend of the characteristic quantity; wherein, the derivation curve DPPMDi is the first derivative curve of the maximum power spectral density;

Step 5: Find the inflection point of the DPPMDi curve, and define the first inflection point as the node where the bearing enters the normal working state, and the second inflection point as the node where the bearing enters the performance decline stage, and determine the predictable interval.

2 Experimental process

According to a method for judging the degradation node of rolling bearing performance degradation proposed in the present invention, the bearing degradation data is collected by using the experimental platform of the PRONOSTIA laboratory of the FEMTO-ST research institute. Bearing degradation data are collected using vibration acceleration sensors placed at 90° to each other. The first sensor is placed in the vertical direction of the bearing, the second sensor is placed in the horizontal direction of the bearing, and the two sensors are placed radially on the outer race of the bearing. , the sampling frequency is 25.6kHz, and 2560 data samples are collected every 10 seconds. The experimental data used in the present invention is the data collected by the sensor in the vertical direction.

The three operating conditions of the experiment are: (1) 1800rpm and 4000N; (2) 1650rpm and 4200N;

(3) 1500rpm and 5000N.

Bearings are subjected to life experiments under different operating conditions, and the specific distribution is shown in Table 1.

Table 1 Bearing life dataset

Taking bearing1_1 as an example, the collected vibration acceleration data of the first monitoring period are shown in Table 2.

Table 2 The vibration acceleration data of the first monitoring cycle of bearing1_1

The power spectral density is obtained after Hilbert transform is performed on the vibration acceleration data of each monitoring period.

Taking bearing1_1 as an example, the power spectral density value of the vibration acceleration data collected in the first monitoring period is shown in Table 3.

Table 3 The power spectral density value of the vibration acceleration data of the first monitoring cycle of bearing1_1

Then, the maximum value operation is performed on the power spectral density of the vibration acceleration data in each monitoring period.

Taking the bearing bearing1_1 as an example, the maximum power spectral density of the vibration acceleration data throughout the life cycle is shown in Table 4.

Table 4 The maximum power spectral density of the vibration acceleration data of the bearing1_1 in the whole life cycle

3 Experimental results and analysis

According to the specific algorithm flow of the performance evaluation model, the maximum power spectral density curve and its 4th order fitting curve, the first derivation curve of the 4th order fitting curve, and the full life cycle characteristic curve of all bearings are made for the full life data of all bearings. As shown in Figure 2, it can be seen from Figure 2 that in the whole life cycle, the Hilbert spectrum of each bearing presents different shapes, and the shape of the four-order fitting curve is also different, but the four-order fitting curve is different. The first derivative of the composite curve shows a similar shape, with two distinct inflection points. It can be seen from this that the new model for performance evaluation of rolling bearings proposed by the present invention can make the change trend of the extracted performance degradation characteristic quantities of different bearings consistent in the whole life cycle, and can determine the working stage of the bearing according to the trend curve. And the node entering the recession period, and then determine the predictable interval.

Comparative analysis of different performance degradation characteristic indicators

In order to verify the validity of the method for determining the degradation node of rolling bearing performance proposed in the present invention, the time domain, higher order mathematical morphology spectral entropy (HOMMSE), root mean square (Root mean square, RMS), energy entropy (Energy entropy, EE), these four performance degradation characteristic indicators are compared with PMD indicators, and the bearing1_1 and bearing1_2 are selected as the analysis objects, and the comparison results are shown in Figure 3.

In this experiment, two bearings and five kinds of bearing life characteristic curves are listed. Figure 3(a) is the time domain diagram of the vibration acceleration of the bearing throughout its life cycle, from which it can be seen that the amplitude of the vibration acceleration of the bearing bearing1_1 gradually increases with time until it is finally damaged; the curve shape of the bearing bearing1_2 is obviously different from that of the bearing1_1. There are a lot of sudden changes on the curve, and there is no obvious law in the amplitude change, indicating that there is a lot of noise in the vibration; Figure 3(b) is the high-order mathematical morphology spectrum entropy curve of the bearing, from which we can see that the bearing From the beginning to the damage of bearing1_1, the entropy value shows an upward trend, while the change of the entropy value of bearing1_2 is not obvious, and the trend change is basically invisible, indicating that the HOMMSE curve is sensitive to noise; Figure 3(c) is the RMS value curve of the bearing From the figure, it can be seen that the RMS value of bearing1_1 gradually increases until the bearing is damaged, while the RMS amplitude of bearing1_2 has obvious mutation, which is also caused by a large amount of noise in the collected signal species; Figure 3(d) is the bearing EE From the graph, we can see that the curve of bearing1_1 is high in the middle and low on both sides, while the curve of bearing1_2 is covered by a large number of sudden changes. Comprehensive time domain, HOMMSE, RMS, EE curves, the experimental results show that these four characteristics are sensitive to noise and cannot shield the influence of noise on the curve; different from the above characteristic curves, Figure 3(e) is the bearing PMD curve, It can be seen that there are obvious similarities in the curves of bearings bearing1_1 and bearing1_2, and their overall changing trends are similar. There are two obvious inflection points, and when the bearing enters the recession period, its value gradually increases until it finally fails.

The above disclosures are only specific embodiments of the present invention, however, the embodiments of the present invention are not limited thereto, and any changes that can be conceived by those skilled in the art should fall within the protection scope of the present invention.

Claims (1)

1. A method for judging a rolling bearing performance degradation node is characterized by comprising the following specific steps:

step 1: obtaining vibration acceleration data of different monitoring periods, and performing Hilbert transform on data samples Dataj in a bearing full-life data set bi to obtain the power spectral density PSD _ Dataj of the Dataj;

step 2: obtaining the maximum value PSD _ MAX _ Dataj of a power spectrum density curve PSD _ Dataj of the data Dataj, and forming a power spectrum density maximum value set PMDI (PSD _ MAX _ Dataj | i is 1,2, …, n, j is 1,2, … and Ni), wherein n is the number of bearings contained in a bearing full-life data set bi, and Ni is the number of data samples in bi;

and step 3: performing curve fitting on the power spectral density maximum value set PMDI by using fourth-order polynomial fitting to obtain a fitting curve PPMDi;

and 4, step 4: performing derivation on the fitting curve PPMDi to obtain a derivation curve DPPMDI of the fitting curve, namely the characteristic quantity change trend; the derivative curve DPPMDi is a first derivative curve of the power spectral density maximum value fitting curve PPMDi;

and 5: and solving inflection points of the DPPMDI curve, defining the first inflection point as a node of the bearing entering a normal working state, defining the second inflection point as a node of the bearing entering a performance degradation stage, and determining a predictable interval.

Images