Time-Series Accelerated Degradation Point Identification Method Based on the Plane Angle of Evaluation Index

Abstract

The present invention discloses a method for identifying accelerated degradation points in a time series based on the plane angle of an evaluation index, which belongs to the field of mechanical equipment status assessment and remaining life prediction. The method collects the horizontal vibration signal of the rolling bearing under test; divides the time series into multiple time units and extracts the absolute peak value within each time unit; uses the cumulative and absolute peak values as rolling bearing status evaluation indicators to characterize the cumulative degradation effect during the operation of the rolling bearing; calculates the angle between the evaluation index and the positive direction of the x-axis of the coordinate system, and traverses each angle value in the angle time series data set; constructs three conditions for identifying the accelerated degradation point of the rolling bearing, and seeks the time point corresponding to the angle data value that simultaneously meets the three conditions. If a certain time point simultaneously meets the above three conditions, the accelerated degradation point of the rolling bearing is identified. The present invention can improve the accuracy of identifying the accelerated degradation point of the rolling bearing, thereby improving the accuracy of predicting the remaining life of the rolling bearing.

Description

Technical Field

[0001] The present invention relates to a time series accelerated degradation point identification method based on evaluation index plane angle, belonging to the field of mechanical equipment state evaluation and remaining life prediction. Background Art

[0002] Rolling bearings are critical components of rotating machinery, and their performance directly impacts the operation of the entire mechanical system. During operation, rolling bearing performance gradually degrades, and the corresponding vibration signature changes as well. Therefore, building performance evaluation metrics based on rolling bearing vibration signatures to assess degradation and, subsequently, predict remaining life based on these metrics to accurately allocate maintenance resources is crucial for improving the reliability of mechanical equipment.

[0003] The degradation process of rolling bearings can be roughly divided into a stable degradation stage and an accelerated degradation stage. The segmentation point between the stable degradation stage and the accelerated degradation stage is called the accelerated degradation point. The remaining life prediction work based on evaluation indicators requires selecting data from the accelerated degradation point of the bearing and then training based on the corresponding prediction model to predict the remaining life. Therefore, the location selection of the accelerated degradation point will affect the accuracy of the remaining life prediction results of the rolling bearing.

[0004] In summary, the field of condition assessment and life prediction urgently needs to solve the problem of identifying accelerated degradation points. Summary of the Invention

[0005] In order to solve the problem that it is difficult to determine the accelerated degradation point of a rolling bearing, the main purpose of the present invention is to provide a time series accelerated degradation point identification method based on the plane angle of the evaluation index, collect the horizontal vibration signal of the tested rolling bearing, and form a time series set of horizontal vibration signals; divide the time series into multiple time units, and extract the absolute peak value within each time unit; then solve and obtain the cumulative and absolute peak values corresponding to the rolling bearing in each time unit, and use the cumulative and absolute peak values as rolling bearing status evaluation indicators to characterize the cumulative degradation effect during the operation of the rolling bearing. Obtain the angle between the evaluation index and the positive direction of the x-axis of the coordinate system to obtain a set of time series data sets about the angle; traverse each angle value in the angle time series data set in turn to find the time point corresponding to the angle data value that satisfies the following three conditions at the same time: Condition ①: the angle data values corresponding to all time points after the time point are greater than or equal to the angle data value corresponding to the time point; Condition ②: the positive growth index of the angle data value corresponding to the short-term stage after the time point is greater than or equal to 0.9; Condition ③: the number of data whose angle data value corresponding to the short-term stage before the time point is greater than the angle data value corresponding to the time point is less than or equal to 10% of the total amount of data in the short-term stage. If a certain time point satisfies the above three conditions at the same time, the accelerated degradation point of the rolling bearing is identified. The present invention can improve the accuracy of identifying the accelerated degradation point of the rolling bearing, thereby improving the accuracy of predicting the remaining life of the rolling bearing.

[0006] The purpose of the present invention is achieved through the following technical solutions:

[0007] The present invention discloses a method for identifying time series degradation points based on the evaluation index plane angle, comprising the following steps:

[0008] Step 1: Collect the horizontal vibration signal of the rolling bearing to form a time series W = [V1, V2, ..., V N ], N is the number of time series samples; the horizontal vibration signal time series is divided, and the sampling signal in each time unit is V=[V1,V2,…,V k ], k is the number of sampling points per unit time, and a total of N / k time units are obtained.

[0009] Step 2: Extract the absolute peak value within each time unit to form an absolute peak time series set.

[0010] Step 3: Determine the cumulative and absolute peak values for each time unit of the rolling bearing under test based on the absolute peak value time series. This cumulative and absolute peak values are used as bearing condition assessment indicators to characterize the cumulative degradation effects of the rolling bearing during operation. The cumulative and absolute peak values for a particular time unit are equal to the sum of the absolute peak values for that time unit and all previous time units.

[0011] Step 4: Use the bearing condition evaluation index constructed in step 3 as input, and calculate the angle between the measured rolling bearing evaluation index and the positive direction of the coordinate system x-axis according to the angle solution formula to obtain a set of time series data sets about the angle.

[0012] Step 5: Construct three conditions for identifying the degradation point of rolling bearings: Condition ①: the angle data values corresponding to all time points after this time point are greater than or equal to the angle data value corresponding to this time point; Condition ②: the positive growth index of the angle data value corresponding to the short-term stage after this time point is greater than or equal to 0.9; ③ the number of data whose angle data value corresponding to the short-term stage before this time point is greater than the angle data value corresponding to this time point is less than or equal to 10% of the total data in the short-term stage. According to the angle time series data set obtained in step 4, each angle value of the angle data set is traversed in turn to determine whether the time point corresponding to the angle value satisfies the three conditions of the accelerated degradation point of the rolling bearing at the same time. If the time point fails to meet the three conditions of the accelerated degradation point of the rolling bearing at the same time, it is determined that the time point is not the accelerated degradation point of the rolling bearing. Continue traversing until the three conditions of the accelerated degradation point of the rolling bearing are met at the same time, then it is determined that the time point is the accelerated degradation point of the rolling bearing, that is, the accelerated degradation point of the rolling bearing is identified based on the time series of the evaluation index plane angle, and the accuracy of the identification of the accelerated degradation point of the rolling bearing is improved based on the identification results, thereby improving the prediction accuracy of the remaining life of the rolling bearing.

[0013] The short-term phase in step 5 refers to the time series process corresponding to the N*P number of data. Preferably, the value range of P is 1.5% to 2%. In order to improve the accuracy of the identification result, as a further preferred embodiment, the value range of P is 1.5%.

[0014] The process of extracting the absolute peak value in each time unit in step 2 is as follows: the sampling signal of the tested bearing in a certain unit time is V i =[V1,V2,…,V k ], where j = 1, 2, ..., k, and k represents the number of sampling points within a unit time point. The absolute peak value within a certain time unit is calculated as follows:

[0015] Peak i =max|v j | (1)

[0016] According to formula (1), the absolute peak values in all time units are obtained, and the time series set of absolute peak values is obtained, which is expressed as [Peak1, Peak2,…, Peak N / k ], i=1,2,…,N / k.

[0017] The cumulative and absolute peaks (CBA) corresponding to a certain time unit in step 3 is equal to the sum of the absolute peaks corresponding to the time unit and all time units before the time unit; the data set of the cumulative and absolute peaks can be expressed as [CBA1, CBA2, ..., CBA N / k ], i = 1, 2, …, N / k, and the cumulative and absolute peak values are used as bearing condition evaluation indicators.

[0018] The method for solving the angle between the evaluation index of the rolling bearing under test and the positive direction of the x-axis of the coordinate system in step 4 is as follows: the order of a certain time unit is located at the x-th position of all time units, and the corresponding evaluation index is y. Then the angle between the evaluation index and the positive direction of the x-axis is solved by the formula:

[0019]

[0020] In formula (2), atan represents the inverse tangent function. According to formula (2), the angle between the evaluation index value and the positive direction of the x-axis at all times is solved, and the time series data set of the angle between the evaluation index value and the positive direction of the x-axis is obtained as [angle1,angle2,…,angle N / k ], i=1,2,…,N / k.

[0021] The implementation method of step 5 includes the following sub-steps:

[0022] Step 5.1: Starting from the angle value corresponding to the first time point in the angle data set in step 4, traverse in sequence to find the required time point. This time point needs to meet condition ①: the angle data values corresponding to all time points after this time point are greater than or equal to the angle data value corresponding to this time point. If the time point meets the above conditions, proceed to step 5.2;

[0023] Step 5.2: Based on the time point obtained in step 5.1, determine whether the time point meets condition ②: the positive growth index of the angle data value corresponding to the short-term period after the time point is greater than or equal to 0.9. If the time point meets condition ②, proceed to step 5.3; otherwise, return to step 5.1. The short-term period refers to the time series process corresponding to N*P data. Preferably, the value range of P is 1.5% to 2%; to improve the accuracy of the identification results, as a further preferred range, the value range of P is 1.5%. The method for solving the positive growth index is as follows:

[0024] Time Series Performing first-order difference on the time series yields a differential time series set, which can be expressed as Then the solution formula for the positive growth index of the time series is:

[0025]

[0026] In formula (3), num(df>0) represents the number of values greater than 0 in the differential time series set, num(df=0) represents the number of values equal to 0 in the differential time series set, num(df<0) represents the number of values equal to 0 in the differential time series set, and num(df) represents the total number of data in the differential time series.

[0027] Step 5.3: Based on the time point obtained in step 5.2, determine whether the time point meets condition ③: the number of data points whose angle data values corresponding to the short-term phase preceding the time point are greater than the angle data values corresponding to the time point is less than or equal to 10% of the total short-term phase data. The short-term phase refers to a time series process corresponding to N*P data points. Preferably, the value range of P is 1.5% to 2%. To improve the accuracy of the identification results, it is further preferred that the value range of P is 1.5%. If the condition is met, the time point is the identified accelerated degradation point, and the operation ends. If not, return to step 5.1 until a time point that meets the conditions described in steps 5.1, 5.2, and 5.3 is found. This point is the accelerated degradation point of the rolling bearing.

[0028] Beneficial effects:

[0029] 1. The present invention discloses a method for identifying accelerated degradation points in a time series based on the plane angle of an evaluation index. The method extracts the absolute peak value within each unit time in the rolling bearing time series, solves the cumulative and absolute peak values corresponding to the tested rolling bearing in each time unit based on the absolute peak time series set, and uses the cumulative and absolute peak values as the bearing state evaluation index. The cumulative and absolute peak values are equal to the sum of the absolute peak values corresponding to the time unit and all time units before the time unit. That is, by accumulating and processing the absolute peak values, an evaluation index reflecting the rolling bearing degradation process is constructed. The angle between the evaluation index and the positive direction of the x-axis in the coordinate system can reflect the performance change trend of the rolling bearing. When the rolling bearing enters the accelerated degradation state from a stable degradation state, the evaluation index will gradually increase, and the corresponding angle data value between the evaluation index and the positive direction of the x-axis will also gradually increase, which is manifested as a phenomenon that the angle data value between the evaluation index and the positive direction of the x-axis at the accelerated degradation point of the bearing will change significantly. The phenomenon is summarized and three conditions for identifying the degradation point of the rolling bearing are proposed. The time point corresponding to the simultaneous satisfaction of the three conditions is the accelerated degradation point of the bearing. The present invention identifies degradation points based on angle data values corresponding to evaluation indicators, and the identification results are more efficient, thereby improving the prediction accuracy of the remaining service life.

[0030] 2. The present invention discloses a method for accelerating degradation point identification based on the plane angle of the evaluation index in a time series, which extracts the absolute peak value in each unit time in the rolling bearing time series, and obtains an evaluation index that can describe the rolling bearing degradation process by accumulating and processing the absolute peak value; unlike the traditional evaluation index root mean square, this index characterizes the cumulative degradation process of the bearing through the cumulative sum approach; during the actual operation of the bearing, the degradation amount of the bearing at the current moment will accumulate and continuously affect the bearing performance at subsequent moments, so the evaluation index constructed in the present invention is more in line with the actual operating conditions of the bearing. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a time series accelerated degradation point identification method based on the evaluation index plane angle of the present invention;

[0032] Figure 2 is the absolute peak time domain diagram of the test bearing Bearing1_1;

[0033] Figure 3 The CBA time domain diagram and RMS time domain diagram of the test bearing Bearing1_1 are shown in the figure. Figure 3 a) is the CBA time domain diagram of the test bearing Bearing1_1, 3b) is the RMS time domain diagram of the test bearing Bearing1_1;

[0034] Figure 4 is the time domain diagram of the evaluation index of each test bearing and the angle in the positive direction of the x-axis under three working conditions, where Figure 4 a) is a time domain diagram of the evaluation index of the test bearing Bearing1_1 and the angle in the positive direction of the x-axis under working condition 1, 4b) is a time domain diagram of the evaluation index of the test bearing Bearing1_2 and the angle in the positive direction of the x-axis under working condition 1, 4c) is a time domain diagram of the evaluation index of the test bearing Bearing2_1 and the angle in the positive direction of the x-axis under working condition 2, 4d) is a time domain diagram of the evaluation index of the test bearing Bearing2_2 and the angle in the positive direction of the x-axis under working condition 2, 4e) is a time domain diagram of the evaluation index of the test bearing Bearing3_1 and the angle in the positive direction of the x-axis under working condition 3, 4f) is a time domain diagram of the evaluation index of the test bearing Bearing3_2 and the angle in the positive direction of the x-axis under working condition 3;

[0035] Figure 5 is the identification result of the accelerated degradation point of the test bearing Bearing1_1 under working condition 1;

[0036] Figure 6 is the identification result of the accelerated degradation point of the test bearing Bearing1_2 under working condition 1;

[0037] Figure 7is the identification result of the accelerated degradation point of the test bearing Bearing2_1 under working condition 2;

[0038] Figure 8 is the identification result of the accelerated degradation point of the test bearing Bearing2_2 under working condition 2;

[0039] Figure 9 is the identification result of the accelerated degradation point of the test bearing Bearing3_1 under working condition 3;

[0040] Figure 10 This is the result of identification of the accelerated degradation point of the test bearing Bearing3_2 under working condition three. DETAILED DESCRIPTION

[0041] In order to better illustrate the purpose and advantages of the present invention, the invention is further described below with reference to the accompanying drawings and examples.

[0042] like Figure 1 As shown; the time series accelerated degradation point identification method based on the evaluation index plane angle disclosed in this embodiment is specifically implemented as follows:

[0043] Step 1: Collect the horizontal vibration signal of the rolling bearing to form a time series W = [V1, V2, ..., V N ], N is the number of time series samples; the horizontal vibration signal time series is divided, and the sampling signal in each time unit is V=[V1,V2,…,V k ], k is the number of sampling points per unit time, and a total of N / k time units are obtained;

[0044] Step 2: Extract the absolute peak value within each time unit to form an absolute peak time series set;

[0045] Step 3: Determine the cumulative and absolute peak values for each time unit of the rolling bearing under test based on the absolute peak time series set. This cumulative and absolute peak values are used as bearing condition assessment indicators to characterize the cumulative degradation effects of the rolling bearing during operation. The cumulative and absolute peak values for a particular time unit are equal to the sum of the absolute peak values for that time unit and all previous time units.

[0046] Step 4: Use the bearing condition evaluation index constructed in step 3 as input and calculate the angle between the rolling bearing evaluation index and the positive direction of the x-axis of the coordinate system according to the angle solution formula to obtain a time series data set of the angle.

[0047] Step 5: Construct three conditions for identifying rolling bearing degradation points: Condition 1: The angle data values corresponding to all time points after the time point are greater than or equal to the angle data value corresponding to the time point; Condition 2: The positive growth index of the angle data values corresponding to the short-term period after the time point is greater than or equal to 0.9; Condition 3: The number of data points whose angle data values corresponding to the short-term period before the time point are greater than the angle data value corresponding to the time point is less than or equal to 10% of the total short-term period data. Based on the angle time series data set obtained in Step 4, traverse each angle value in the angle data set in turn to determine whether the time point corresponding to the angle value simultaneously meets the three conditions for the rolling bearing accelerated degradation point. If the time point does not simultaneously meet the three conditions for the rolling bearing accelerated degradation point, it is determined that the time point is not the rolling bearing accelerated degradation point. Continue traversing until all three conditions for the rolling bearing accelerated degradation point are simultaneously met, and then determine that the time point is the rolling bearing accelerated degradation point.

[0048] The short-term phase in step 5 refers to the time series process corresponding to the N*P number of data. Preferably, the value of P ranges from 1.5% to 2%. In order to improve the accuracy of the identification result, as a further preferred embodiment, the value of P is 1.5%.

[0049] Example

[0050] The data comes from the IEEE PHM 2012 Challenge, provided by the French FEMTO-ST Institute. The bearing vibration signal was sampled at a frequency of 25.6 kHz, with samples taken every 10 seconds for a sampling period of 0.1 seconds. The adjustable operating conditions included speed and radial force. The test bench was configured with three different operating conditions: Condition 1 (speed of 1800 rpm, radial force of 4000 N), Condition 2 (speed of 1650 rpm, radial force of 4200 N), and Condition 3 (speed of 1500 rpm, radial force of 5000 N). Training and test data for each of these conditions were provided. Due to the large number of test data types, the proposed method was implemented using the training data from each condition as an example.

[0051] Step 1: Collect the horizontal vibration signal of the rolling bearing to form a time series W = [V1, V2, ..., V N ], N is the number of time series samples; the horizontal vibration signal time series is divided, and the sampling signal in each time unit is V=[V1,V2,…,V k ], k is the number of sampling points per unit time, and a total of N / k time units are obtained;

[0052] Taking the bearing Bearing1_1 under the first working condition as an example, the horizontal vibration signal of the rolling bearing under test is collected to form a time series W = [V1, V2, ..., V N ], N is the number of sequence samples, N = 7175680; the horizontal vibration signal time series is divided, and the sampling signal in each time unit is V = [V1, V2, ..., V k ], k is the number of sampling points per unit time, k = 2560; the time series of the horizontal vibration signal can be divided into 2803 time units in total.

[0053] Step 2: Extract the absolute peak value within each time unit to form an absolute peak time series set.

[0054] The sampling signal of the tested bearing in a certain time unit is V i =[V1,V2,…,V k ], where j = 1, 2, ..., k, and k represents the number of sampling points in a time unit. The absolute peak value calculation formula in a time unit is as follows:

[0055] Peak i =max|v j

[0056] Taking the bearing Bearing1_1 under the first working condition as an example, the time domain diagram of the absolute peak value within each time unit is extracted as follows: Figure 2 shown.

[0057] Step 3: Calculate the cumulative and absolute peak values corresponding to each time unit of the tested rolling bearing based on the absolute peak time series set, and use the cumulative and absolute peak values as bearing condition evaluation indicators to characterize the cumulative degradation effect of the rolling bearing during operation.

[0058] The cumulative and absolute peaks (CBA) of the tested rolling bearing corresponding to a certain time unit is equal to the sum of the absolute peaks corresponding to the time unit and all time units before the time unit; the data set of the cumulative and absolute peaks can be expressed as [CBA1, CBA2,…, CBA N / k ], i = 1, 2, …, N / k, and the cumulative and absolute peak values are used as bearing condition evaluation indicators.

[0059] Taking the bearing Bearing1_1 under the first working condition as an example, the evaluation index is solved. In order to compare the superiority of the evaluation index, the traditional evaluation index root mean square (RMS) is obtained. Then the time domain diagram of the evaluation index (CBA) constructed by this patent and the time domain diagram of the root mean square are plotted on Figure 3 .

[0060] like Figure 3 As shown, no matter what working conditions the training bearing is in, the curve change trend of the evaluation index CBA proposed by the present invention has a very obvious upward trend with the increase of operating time. This is consistent with the actual operating conditions of the bearing. The degradation of the bearing is mainly caused by wear or deformation, and the wear or deformation at the current moment will have a certain impact on the state of the bearing at subsequent moments. Therefore, the degradation of the bearing is a cumulative degradation process. The evaluation index of the bearing at each moment should reflect the cumulative degradation of the bearing performance. The evaluation index constructed by the present invention can just reflect the above-mentioned change process. However, observing the time domain diagram of the traditional evaluation index RMS, it can be seen that RMS cannot describe the cumulative degradation process of the bearing. Therefore, the bearing evaluation index constructed by the present invention has better performance.

[0061] Monotonicity is an important indicator for evaluating bearing performance. To demonstrate the advantages of the evaluation indicators constructed by the present invention, the monotonicity of CBA and RMS of each test bearing under different working conditions is calculated. The monotonicity calculation process is as follows:

[0062] A set of time series is [x1,x2,…,x N / k ], the first-order difference of the time series is obtained to obtain the differential time series set, which can be expressed as The formula for solving the monotonicity of time series is:

[0063]

[0064] Where, num(df x >0) represents the number of differential time series that is greater than 0, num(df x <0) represents the number of zeros in the differential time series, and num(df) represents the number of data points in the differential time series. The monotonicity of the CBA and RMS of each test bearing under different operating conditions was then calculated, as shown in Table 1.

[0065] Table 1 Monotonicity of CBA and RMS of tested bearings

[0066]

[0067] As shown in Table 1, the monotonicity of the evaluation metric CBA constructed in the present invention is 1 under all operating conditions, significantly exceeding the monotonicity of the traditional evaluation metric RMS. This further demonstrates the superiority of the evaluation metric constructed in the present invention in describing the bearing degradation process.

[0068] Step 4: Use the bearing condition evaluation index constructed in step 3 as input and calculate the angle between the rolling bearing evaluation index and the positive direction of the x-axis of the coordinate system according to the angle solution formula to obtain a time series data set of the angle.

[0069] The method for obtaining the angle between the evaluation index corresponding to each time unit of the tested rolling bearing and the positive direction of the x-axis of the coordinate system is as follows. The order of a certain time unit is located at the x-th position of all time units, and the corresponding evaluation index is y. Then the angle between the evaluation index and the positive direction of the x-axis is solved as follows:

[0070]

[0071] In the formula, atan represents the inverse tangent function, and then solve the angle between the evaluation index value and the x-axis at all times, and get the data set of the angle between the evaluation index value and the x-axis at all times as [angle1,angle2,…,angle N / k ], i=1,2,…,N / k.

[0072] Solve the angle data sets of each test bearing under three working conditions respectively, and draw the time domain diagram as shown in the following figure: Figure 4 shown.

[0073] like Figure 4 As shown, the angle time-domain plot for each test bearing operating condition can be roughly divided into two parts. The first part shows the angle data fluctuating from large fluctuations to a stable state. This is mainly due to the unstable performance at the beginning of the operation of the bearing, which causes large fluctuations in the evaluation index. After a period of operation, it gradually stabilizes and enters the stable degradation stage. The second part shows the angle data gradually increasing. This is mainly due to the accelerated degradation stage of the bearing, with gradually declining performance, resulting in increasingly large fluctuations in the evaluation index. This phenomenon is reflected in the increasing angle data in the time domain. Based on a thorough analysis of the curve trend of the angle data set, step 5 is carried out.

[0074] Step 5: Construct three conditions for identifying rolling bearing degradation points: Condition 1: The angle data values corresponding to all time points after the time point are greater than or equal to the angle data value corresponding to the time point; Condition 2: The positive growth index of the angle data values corresponding to the short-term period after the time point is greater than or equal to 0.9; 3: The number of data points whose angle data values corresponding to the short-term period before the time point are greater than the angle data value corresponding to the time point is less than or equal to 10% of the total short-term period data. Based on the angle time series data set obtained in Step 4, traverse each angle value in the angle data set in turn to determine whether the time point corresponding to the angle value simultaneously meets the three conditions for the rolling bearing accelerated degradation point. If the time point does not simultaneously meet the three conditions for the rolling bearing accelerated degradation point, it is determined that the time point is not the rolling bearing accelerated degradation point. Continue traversing until the three conditions for the rolling bearing accelerated degradation point are simultaneously met, and then determine that the time point is the rolling bearing accelerated degradation point.

[0075] The short-term phase in step 5 refers to the time series process corresponding to the N*P number of data. Preferably, the value of P ranges from 1.5% to 2%. In order to improve the accuracy of the identification result, as a further preferred embodiment, the value of P is 1.5%.

[0076] Step 5 also includes the following sub-steps:

[0077] Step 5.1: Starting from the angle value corresponding to the first moment point of the angle data set in step 4, traverse in sequence to find the required moment point. The moment point needs to meet condition ①: the angle data values corresponding to all moment points after the moment point are greater than or equal to the angle data value corresponding to the moment point. If a moment point that meets the above conditions is obtained, proceed to step 5.2.

[0078] Since the angle data gradually increases after the accelerated degradation point, the subsequent angle data must be larger than the angle corresponding to the accelerated degradation point. Step 5.1 requires the accelerated degradation point to meet the above analysis requirements.

[0079] Step 5.2: Based on the time point obtained in step 5.1, determine whether the time point meets condition ②: the positive growth index of the angle data value corresponding to the short-term period after the time point is greater than or equal to 0.9. If the time point meets condition ②, continue to step 5.3; otherwise, return to step 5.1. The short-term period refers to the time series process corresponding to N*P data. The value of P is 1.5%. The method for solving the positive growth index is as follows:

[0080] The time series is [angle'1,angle'2,…,angle' N / k ], the first-order difference of the time series is obtained to obtain the differential time series set, which can be expressed as Then the solution formula for the positive growth index of the time series is:

[0081]

[0082] Where num(df>0) represents the number of values greater than 0 in the differential time series set, num(df=0) represents the number of values equal to 0 in the differential time series set, num(df<0) represents the number of values equal to 0 in the differential time series set, and num(df) represents the number of data in the differential time series.

[0083] After the accelerated degradation point, the value of the angle data in the short term should maintain a positive growth trend relative to the angle value at the accelerated degradation point. However, due to the influence of random factors in actual situations, there may be a situation where the angle value at the accelerated degradation point increases in the short term and then a small number of angle values show a downward trend. The acceptance level of this situation is defined as the positive growth index of the angle data at the accelerated degradation point in the short term is greater than or equal to 0.9.

[0084] Step 5.3: Based on the time point obtained in step 5.2, determine whether the time point meets condition ③: the number of data points whose angle data values corresponding to the short-term phase preceding the time point are greater than the angle data values corresponding to the time point is less than or equal to 10% of the total short-term phase data. The short-term phase refers to a time series process corresponding to N*P data points. Preferably, the value range of P is 1.5% to 2%. To improve the accuracy of the identification results, it is further preferred that the value range of P is 1.5%. If the condition is met, the time point is the identified accelerated degradation point, and the operation ends. If not, return to step 5.1 until a time point that meets the conditions described in steps 5.1, 5.2, and 5.3 is found. This point is the accelerated degradation point of the rolling bearing.

[0085] As the bearing degrades, the angle gradually increases. As the bearing enters the accelerated degradation stage from the stable degradation stage, the angle data also maintains an increasing trend. Therefore, in the short period before the accelerated degradation point, the angle data should be smaller than the angle corresponding to the accelerated degradation point. However, due to random influences such as environmental factors, the angle may occasionally increase in the early stage of the accelerated degradation point. The values caused by these random factors are required to be less than or equal to 10% of the total data in the short period. Therefore, this step ensures that the above analysis requirements are met.

[0086] The proposed method was then used to determine the degradation points for each test bearing under different operating conditions. To demonstrate the superiority of the proposed method, a degradation point identification method based on the traditional RMS (root mean square) evaluation metric, Three Sigma, was presented. The final comparison results of the accelerated degradation point identification are shown in Table 2.

[0087] Table 2 Accelerated degradation point identification results

[0088]

[0089] Since there is no objective comparison object for the accelerated degradation point of the test bearing, and there is no relevant theoretical formula to determine it, the change in the vibration signal is the most intuitive reflection of the bearing performance. Therefore, the change in the bearing vibration signal can be used to roughly identify the location of the accelerated degradation point. Generally speaking, at the accelerated degradation point, the vibration signal of the bearing will have a more obvious increase. According to the results in Table 2, the accelerated degradation points identified by different methods are marked on the vibration signal of the test bearing, and a comparative analysis is carried out. The identification results are as follows: Figures 5 to 10 shown.

[0090] like Figure 5 As shown in the figure, it can be seen that the vibration signal of the test bearing Bearing1_1 gradually increases near the time point of 1300. The accelerated degradation point identified by the present invention is at the time point of 1345, which is basically consistent with the change of the vibration signal. However, the three sigma method identifies the accelerated degradation point at the time point of 25, which is inconsistent with the change of the vibration signal.

[0091] like Figure 6 As shown in the figure, it can be seen that the vibration signal of the test bearing Bearing1_2 gradually increases near the time point 826. The degradation point positions identified by the method proposed in the present invention and the three-sigma method are almost the same, but the method proposed in the present invention is one time point earlier than the three-sigma method.

[0092] like Figure 7 As shown in the figure, it can be seen that the vibration signal of the test bearing Bearing2_1 gradually increases near the time point of 150. The accelerated degradation point identified by the present invention is at the time point of 147, which is basically consistent with the change of the vibration signal. The three sigma method identifies the accelerated degradation point at the time point of 34. Figure 7 It can be seen from the figure that the vibration signal does increase suddenly at time 34, but after this point the vibration signal of the bearing tends to be stable. Therefore, time 34 is a random situation and cannot be determined as an accelerated degradation point.

[0093] like Figure 8 As shown in the figure, it can be seen that the vibration signal of the test bearing Bearing2_2 gradually increases around the time point 200. The accelerated degradation point identified by the present invention is at the time point 204, while the three sigma method only detects the increase in the vibration signal at the time point 230. The method proposed in the present invention can monitor the vibration signal in time, 46 time points earlier than the three sigma method.

[0094] like Figure 9As shown in the figure, it can be seen that the vibration signal of the test bearing Bearing3_1 gradually increases near the 490 time point. The accelerated degradation point identified by the present invention is at the 490 time point, which is basically consistent with the change of the vibration signal. The three sigma method identifies the accelerated degradation point at the 23 time point. Figure 9 It can be seen from the figure that the vibration signal does increase suddenly at time 23, but after this point the vibration signal of the bearing tends to be stable. Therefore, time 34 is a random situation and cannot be determined as an accelerated degradation point.

[0095] like Figure 10 As shown in the figure, it can be seen that the vibration signal of the test bearing Bearing3_2 gradually increases around the time point of 1400. The accelerated degradation point identified by the present invention is at the time point of 1432, while the three-sigma method only detects the increase in the vibration signal at the time point of 1598. The method proposed in the present invention can monitor the vibration signal in time, 166 time points earlier than the three-sigma method.

[0096] In summary, regardless of the working conditions, the method proposed in the present invention can identify the accelerated degradation point earlier, is more universal and superior than the three sigma method, and is an effective method for identifying the accelerated degradation point.

[0097] The above disclosed specific description further elaborates on the purpose, technical solutions and effective effects of the invention, but the embodiments of the present invention are not limited thereto. Any modifications made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A time series accelerated degradation point identification method based on the evaluation index plane angle, characterized by: The following steps are included: Step 1: Collect the horizontal vibration signal of the rolling bearing to form a time series W = [V1, V2, ..., V N ], N is the number of time series samples; the horizontal vibration signal time series is divided, and the sampling signal in each time unit is V=[V1,V2,…,V k ], k is the number of sampling points per unit time, and a total of N / k time units are obtained; Step 2: Extract the absolute peak value within each time unit to form an absolute peak time series set; Step 3: Calculate the cumulative and absolute peak values corresponding to each time unit of the tested rolling bearing based on the absolute peak time series set, and use the cumulative and absolute peak values as bearing condition evaluation indicators to characterize the cumulative degradation effect of the rolling bearing during operation. The cumulative and absolute peak values corresponding to a certain time unit are equal to the sum of the absolute peak values corresponding to the time unit and all time units before the time unit; Step 4: Use the bearing condition evaluation index constructed in step 3 as input and calculate the angle between the rolling bearing evaluation index and the positive direction of the x-axis of the coordinate system according to the angle solution formula to obtain a time series data set of the angle. Step 5: Construct three conditions for identifying the degradation point of rolling bearings: Condition ①: the angle data values corresponding to all time points after this time point are greater than or equal to the angle data value corresponding to this time point; Condition ②: the positive growth index of the angle data value corresponding to the short-term stage after this time point is greater than or equal to 0.9; ③ the number of data whose angle data value corresponding to the short-term stage before this time point is greater than the angle data value corresponding to this time point is less than or equal to 10% of the total data in the short-term stage; according to the angle time series data set obtained in step 4, traverse each angle value of the angle data set in turn to determine Whether the moment point corresponding to the angle value satisfies the three conditions of the rolling bearing accelerated degradation point at the same time; if the moment point fails to satisfy the three conditions of the rolling bearing accelerated degradation point at the same time, it is determined that the moment point is not the rolling bearing accelerated degradation point, and the traversal is continued until the three conditions of the rolling bearing accelerated degradation point are satisfied at the same time, then the moment point is determined to be the rolling bearing accelerated degradation point, that is, the rolling bearing accelerated degradation point is identified based on the time series of the evaluation index plane angle, and the accuracy of the identification of the rolling bearing accelerated degradation point is improved based on the identification result, thereby improving the prediction accuracy of the remaining life of the rolling bearing; The short-term stage refers to the time series process corresponding to N*P number of data.

2. The time series accelerated degradation point identification method based on the evaluation index plane angle according to claim 1 is characterized in that: The value range of P in step 5 is 1.5% to 2%.

3. The time series accelerated degradation point identification method based on the evaluation index plane angle according to claim 2, characterized in that: The value range of P is 1.5%.

4. The time series accelerated degradation point identification method based on the evaluation index plane angle according to claim 2 or 3, characterized in that: The process of extracting the absolute peak value in each time unit in step 2 is as follows: the sampling signal of the tested bearing in a certain unit time is V i =[V1,V2,…,V k ], where j = 1, 2, ..., k, and k represents the number of sampling points within a unit time point. The absolute peak value calculation formula within a certain time unit is as follows: Peak i =max|v j | (1) According to formula (1), the absolute peak values in all time units are obtained, and the time series set of absolute peak values is obtained, which is expressed as [Peak1, Peak2,…, Peak N / k ], i=1,2,…,N / k.

5. The time series accelerated degradation point identification method based on the evaluation index plane angle according to claim 4 is characterized in that: The cumulative and absolute peak value CBA corresponding to a certain time unit in step 3 is equal to the sum of the absolute peak values corresponding to the time unit and all time units before the time unit; the data set of the cumulative and absolute peak values is expressed as [CBA1, CBA2, ..., CBA N / k ], i = 1, 2, …, N / k, and the cumulative and absolute peak values are used as bearing condition evaluation indicators.

6. The time series accelerated degradation point identification method based on evaluation index plane angle according to claim 5, characterized in that: The method for solving the angle between the evaluation index of the rolling bearing under test and the positive direction of the x-axis of the coordinate system in step 4 is as follows: the order of a certain time unit is located at the x-th position of all time units, and the corresponding evaluation index is y. Then the angle between the evaluation index and the positive direction of the x-axis is solved by the formula: In formula (2), atan represents the inverse tangent function. According to formula (2), the angle between the evaluation index value and the positive direction of the x-axis at all times is solved, and the time series data set of the angle between the evaluation index value and the positive direction of the x-axis is obtained as [angle1,angle2,…,angle N / k ], i=1,2,…,N / k.

7. The time series accelerated degradation point identification method based on evaluation index plane angle according to claim 6, characterized in that: The implementation method of step 5 includes the following sub-steps: Step 5.1: Starting from the angle value corresponding to the first time point in the angle data set in step 4, traverse in sequence to find the required time point. This time point needs to meet condition ①: the angle data values corresponding to all time points after this time point are greater than or equal to the angle data value corresponding to this time point. If the time point meets the above conditions, proceed to step 5.2; Step 5.2: Based on the time point obtained in step 5.1, determine whether the time point meets condition ②: the positive growth index of the angle data value corresponding to the short-term period after the time point is greater than or equal to 0.

9. If the time point meets condition ②, continue to step 5.3; otherwise, return to step 5.

1. The short-term period refers to the time series process corresponding to the N*P number of data; The method for solving the positive growth indicator is as follows: Time series [angle1',angle'2,…,angle' N / k ], the first-order difference of the time series is obtained to obtain the differential time series set, which can be expressed as Then the solution formula for the positive growth index of the time series is: In formula (3), num(df>0) represents the number of values greater than 0 in the differential time series set, num(df=0) represents the number of values equal to 0 in the differential time series set, num(df<0) represents the number of values equal to 0 in the differential time series set, and num(df) represents the total number of data in the differential time series; Step 5.3: Based on the moment point obtained in step 5.2, determine whether the moment point meets condition ③: the angle data value corresponding to the short-term stage before the moment point is greater than the number of data corresponding to the angle data value at the moment point is less than or equal to 10% of the total data in the short-term stage; if the condition is met, the moment point is the identified accelerated degradation point and the operation is terminated. If not, return to step 5.1 until a moment point that meets the conditions described in steps 5.1, 5.2 and 5.3 is found. This point is the accelerated degradation point of the rolling bearing.

Images