Desulfurization pump main shaft bearing service performance evaluation method

Abstract

The invention relates to the technical field of bearings, in particular to a service performance evaluation method for a desulfurization pump main shaft bearing. According to the method, on the basis of the robustness of the median of the data sequence and the unsteady feature of the average value, the mean value of the median and the average value of the bearing vibration data at different time stages represents the service performance of the bearing at the time stage; finding out the minimum mean value of the median and the mean value of the bearing vibration data at different time stages as the vibration local true value of the bearing; the percentage of the ratio of the difference between the service performance of the bearing and the vibration local truth value in different time stages to the vibration local truth value is calculated, the percentage is the variation characteristic of the bearing in different time stages, and therefore the service performance of the desulfurization pump main shaft bearing is evaluated. By analyzing the vibration of the fault bearing, the result obtained by the method is consistent with the standard of the rolling bearing, and the service performance of the bearing can be evaluated; the vibration of the desulfurization pump bearing is analyzed, and the service performance of the desulfurization pump bearing can be accurately reflected.

Description

Technical Field

[0001] This invention relates to the field of bearing technology, specifically a method for evaluating the service performance of a desulfurization pump spindle bearing. Background Technology

[0002] Common failure modes of high-power desulfurization pumps include main shaft fracture, bearing cage wear, plastic deformation of inner and outer rings, roller spalling, and uneven blade wear. Among these, direct main shaft bearing failure accounts for about 65%, and failures related to the main shaft bearing account for 85-90%. It can be seen that the main shaft bearing is the key point for performance analysis and fault diagnosis of high-power desulfurization pumps.

[0003] The main shaft bearing of a high-power desulfurization pump is a core component, playing a crucial role in the pump's safety and reliability. Its service performance exhibits significant uncertainties. During pump operation, the main shaft bearing bears axial and radial loads from the unpredictable performance of the curved blades, reflecting pump body vibration and the rotational stability of the main shaft. Simultaneously, during operation, the pump blades are subjected to scouring by lime water, corrosion from alkaline liquids and acidic gases, temperature fluctuations, and solid impacts generated during reactions, resulting in axial and radial loads of uncertain magnitude and direction. These uncertain loads ultimately rest on the rolling bearings, making their lifespan unpredictable.

[0004] The uncertainty in the lifespan of the main shaft bearings of high-power desulfurization pumps stems from three aspects: firstly, the uncertainty of load magnitude changes; secondly, the uncertainty of load direction changes; and thirdly, the coupling effect of load magnitude and direction changes. The combined effect of these three factors results in a significant variation in the lifespan of the main shaft bearings of high-power desulfurization pumps, ranging from three months to three years. In other words, under the same operating conditions, using the same materials, processes, equipment, and personnel for assembly, the lifespan can differ by approximately tenfold, from a few months to several years. This leads to some main shaft bearings continuing to operate normally well well past their design lifespan, while others fail before reaching their expected lifespan.

[0005] The uncertainty in the performance and service life of the main shaft bearing of the desulfurization pump is manifested in the diversity, complexity and chaos of comprehensive performance indicators such as vibration, friction torque, temperature rise and noise, which makes it very difficult to predict the performance and maintain the high-power desulfurization pump.

[0006] The main problems existing in current research on bearing performance regarding key technologies of desulfurization pump spindle bearings are as follows:

[0007] (1) Problems with the distribution characteristics of the main shaft bearings of high-power desulfurization pumps

[0008] The main shaft bearings of high-power desulfurization pumps are a combination of tapered and cylindrical rolling bearings. During operation, influenced by various factors such as manufacturing, installation, commissioning, and the environment, their vibration, frictional torque, and temperature distribution characteristics exhibit complexity and diversity, posing significant challenges to bearing performance evaluation and prediction. Existing methods struggle to analyze the overall patterns of unknown distribution characteristics in desulfurization pump main shaft bearings. This is because most existing studies focus on the influence of a single bearing factor on bearing performance, failing to adequately reflect the bearing performance under the influence of multiple factors during actual service. Some methods consider the influence of coupling factors, but require knowledge of the data distribution type; the data distribution during bearing service is difficult to predict. Some methods can consider unknown characteristics of the data, but cannot determine the significance level of the data.

[0009] (2) The issue of defining the performance degradation state of the desulfurization pump main shaft bearing

[0010] The main factors affecting the performance of the desulfurization pump main shaft bearing include changing environment, load, lubrication and sealing, and blade wear. Among them, the unstable changes in load and blade wear have obvious mutual coupling effects, and the environment, lubrication and sealing have coupling effects, which makes the distribution and changing trend of the performance parameters of the desulfurization pump main shaft bearing exhibit nonlinear and chaotic characteristics. It is difficult to define the performance degradation state of the desulfurization pump main shaft bearing in practice based solely on the performance characteristics of rolling bearing parameters.

[0011] (3) Relationship between the accuracy ( / mass) loss and failure criteria of the desulfurization pump main shaft bearing

[0012] Most methods consider the impact of various factors and coupling effects on bearing accuracy / quality, but fail to consider the relationship between bearing accuracy / quality loss and bearing failure criteria, and do not establish an equivalence relationship between the two, thus failing to make a correct and effective assessment of the performance of bearings in service. These methods are based on fault diagnosis theory, large-scale data analysis, Gaussian distribution and statistical theory, but lack overall inference, hypothesis testing, analysis of variable uncertainty and function polyvariability, and assessment of nonlinear dynamic characteristic variations; they also fail to consider the problem of unknown distribution characteristics and lack of information in the bearing performance process. Summary of the Invention

[0013] In response to the problems pointed out in the background art, the purpose of this invention is to provide a method for evaluating the service performance of desulfurization pump main shaft bearings. Based on robust data theory, this method integrates median and Huber M estimation methods to robustly process the data, establishes a data measurement system, and evaluates data characteristics. Through the method provided by this invention, it is expected to improve the service efficiency of desulfurization pump main shaft bearings, reduce maintenance costs, and avoid major accidents.

[0014] To achieve the above objectives, the present invention adopts the following technical solution:

[0015] A method for evaluating the service performance of a desulfurization pump spindle bearing includes the following steps:

[0016] S1. Divide the service life of the desulfurization pump main shaft bearing into m time stages, collect vibration data of the desulfurization pump main shaft bearing at different time stages during service, and form a data sequence of m time stages.

[0017] S2. For the vibration data at each time stage, sort them in ascending order to obtain the absolute value sorted sequence;

[0018] S3. According to statistics, find the median of each absolute value sorted sequence;

[0019] S4. Based on the Huber M estimation principle, obtain the improved data sequence for each absolute value sorted sequence;

[0020] S5. Based on statistics, obtain the average value of each improved data sequence;

[0021] S6. Obtain the absolute difference between the mean of each improved data sequence and the median of the corresponding absolute value sorted sequence;

[0022] S7. For the vibration data at each time stage, obtain the local eigenrangement;

[0023] S8. Based on the local eigeninterval, calculate the local true value of the bearing vibration at each time stage;

[0024] S9. Calculate the service coefficient of the bearing at each time stage; the larger the service coefficient value, the worse the service performance.

[0025] In step S1, let Xi be the data sequence of the i-th time stage of the desulfurization pump main shaft bearing during service. The mathematical expression of Xi is as follows:

[0026] (1)

[0027] In equation (1), Xi is the vibration data sequence of the desulfurization pump main shaft bearing in the i-th time stage, i is the time stage number, n is the data number, xi(n) is the n-th data in the i-th time stage, and N is the number of the n-th data in the i-th time stage. The number of data points obtained in each time period, where m is the number of time periods.

[0028] In step S2, the vibration data for the i-th time stage are sorted in ascending order, and the absolute value sorting sequence Yi is expressed as follows:

[0029] (2)

[0030] In equation (2), Yi is the absolute value sorting sequence, i is the time stage number, n is the data number, N is the number of data obtained in the i-th time stage, m is the number of time stages, and yi(n) is the n-th data in the absolute value sorting sequence.

[0031] In step S3, the median of the absolute value sorting sequence of vibration data at the i-th time stage is specifically described as follows:

[0032] (3)

[0033] In equation (3), βi is the median of the absolute value sorting sequence; i is the time stage number; N is the number of data obtained in the i-th time stage; m is the number of time stages; yi(n) is the n-th data in the absolute value sorting sequence; n is the data number; n=1,2,…,N; N is the number of data obtained in the i-th time stage.

[0034] In step S4, the improved data sequence of any absolute value sorted sequence is obtained using the following method:

[0035] Suppose yi(b) and yi(e) are the nth elements in the absolute value sorted sequence, respectively. The data and the first One data point, and Let be two data points from 1, 2, ..., N, where yi(b) ≤ βi and βi ≤ yi(e). Define the sorted sequence yi(b), ..., βi in ascending order as the left sequence; the number of data points in the left sequence is n1; yi(b) is the first data point of the left sequence. Define the sorted sequence βi, ..., yi(e) in ascending order as the right sequence; the number of data points in the right sequence is n2; yi(e) is the last data point of the right sequence.

[0036] According to the Huber M estimation principle, when yi(n) ≤ yi(b), yi(b) replaces yi(n); when yi(n) ≥ yi(e), yi(e) replaces yi(n); thus, the improved data sequence is obtained:

[0037] (4)

[0038] In equation (4), Zi(n1,n2) is the improved data sequence, zi(n;n1,n2) is the nth data in the improved data sequence, i is the time stage number, n is the data number, N is the number of data obtained in the i-th time stage, m is the number of time stages, n1 is the number of data in the left sequence, and n2 is the number of data in the right sequence.

[0039] In step S5, the average value of any improved data sequence is as follows:

[0040] ; (5)

[0041] In equation (5), ηi(n1,n2) is the average value of the improved data sequence, zi(n;n1,n2) is the nth data in the improved data sequence, i is the time stage number, n is the data number, N is the number of data obtained in the i-th time stage, m is the number of time stages, n1 is the number of data in the left sequence, and n2 is the number of data in the right sequence.

[0042] In step S6, the absolute difference between the mean of any improved data sequence and the median of the corresponding absolute value sorted sequence is as follows:

[0043] ; (6)

[0044] In equation (6), Di(n1,n2) is the absolute difference between the average of the improved data sequence and the median of the corresponding absolute value sorted sequence, βi is the median of the absolute value sorted sequence, ηi(n1,n2) is the average of the improved data sequence, i is the time stage number, m is the time stage number, n1 is the number of data in the left sequence, and n2 is the number of data in the right sequence.

[0045] In step S7, the specific method for obtaining the local intrinsic interval [Ki1, Ki2] for the vibration data at the i-th time stage is as follows:

[0046] Based on the robustness of the median in statistics, when N is even, n1 = n2 = 1, 2, ..., N / 2; when N is odd, n1 = n2 = 1, 2, ..., (N+1) / 2; N is the number of data points obtained in the i-th time period; i is the time period number; n1 is the number of data points in the left sequence; n2 is the number of data points in the right sequence.

[0047] By taking different values ​​of n1 and n2, we can obtain the absolute difference Di(n1,n2) between the mean of the improved data sequence and the median of the absolute value sorted sequence.

[0048] According to the robustness theory of modern statistics, for robust data, the significance level is α = (n1 + n2) / N, α ≤ 0.1, with a limit of 0.1. Within the range of 0 to 0.1, the significance level α corresponding to the mean closest to the median in robust data is the significance level of the data. Therefore, for different Di(n1, n2), the corresponding significance level α is calculated, and the Di(n1, n2) with the smallest significance level α is found, which is denoted as Dimin. The first data yi(b) of the left sequence and the last data yi(e) of the right sequence corresponding to Dimin are Ki1 and Ki2, respectively. The local intrinsic interval of the i-th time stage is [Ki1, Ki2], where Ki1 is the lower bound of the local intrinsic interval and Ki2 is the upper bound of the local intrinsic interval.

[0049] In step S8, the mathematical expression for the local true value Ti of the bearing vibration at the i-th time stage is as follows:

[0050] (7)

[0051] In step S9, the mathematical expression for the service coefficient of the bearing in the i-th time stage is as follows:

[0052] (8)

[0053] In equation (8): The minimum value of Ti is given, where Ti is the local true value of vibration at the i-th time stage; Wi is the service coefficient of the bearing at the i-th time stage. The larger the value of Wi, the worse the service performance of the bearing.

[0054] The beneficial effects of this invention are as follows: The method provided by this invention does not require knowledge of the distribution type and confidence level of the data sequence to analyze time series data; the service coefficient integrates the robustness of the median and the variability of the mean, improving the reliability and accuracy of the evaluation results, laying a theoretical foundation for the performance analysis of new desulfurization pump products, and providing different exploration directions for the lack of information theory of unknown distribution time series. Attached Figure Description

[0055] Figure 1 This is a flowchart of the present invention.

[0056] Figure 2 The image shows the bearing vibration curve during the first time stage of Example 1 (when the groove damage diameter is 0 mm).

[0057] Figure 3 The bearing vibration curve is shown in the second time stage of Example 1 (when the groove damage diameter is 0.1778 mm).

[0058] Figure 4This is a bearing vibration curve for the third time stage of Example 1 (when the groove damage diameter is 0.5334 mm).

[0059] Figure 5 This is a bearing vibration curve for the fourth time stage of Example 1 (when the groove damage diameter is 0.7112 mm).

[0060] Figure 6 The median and average values ​​of the rolling bearing vibration robustness data for the first to fourth time periods of Example 1 are given.

[0061] Figure 7 The values ​​represent the local true values ​​of the bearing vibration at different time stages in Example 1.

[0062] Figure 8 The service coefficients for bearing vibration at different time stages in Example 1 are shown.

[0063] Figure 9 The image shows the vibration curve of the 7008AC / P2 rolling bearing in Example 2. Detailed Implementation

[0064] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0065] Example 1

[0066] The experimental data for Example 1 comes from the bearing data center website of Case Western Reserve University in the United States, which has a dedicated rolling bearing failure simulation test bench. The test bench consists of a motor, torque sensor / decoder, and power meter. The SKF6205 rolling bearing under test supports the rotating shaft of the motor. The vibration acceleration of the rolling bearing is measured using an accelerometer, in V. The bearing speed is 1797 r / min, the sampling frequency is 12 kHz, and the damage diameters di of the inner ring groove of the rolling bearing are d1=0 mm, d2=0.1778 mm, d3=0.5334 mm, and d4=0.7112 mm, respectively; i is the time stage number; i=1,2,3,4.

[0067] In this embodiment 1, the rolling bearing vibration data sequences obtained under four damage diameters d1=0mm, d2=0.1778mm, d3=0.5334mm and d4=0.7112mm are simulated as rolling bearing vibration data sequences X1, X2, X3 and X4 obtained in four time stages; i=1,2,3,4; i is the time stage number; the number of time stages m=4; each vibration data sequence has N=1600 data points.

[0068] Since the balls in a rolling bearing wear down gradually, it is possible to... 0 mm,0.1778 Vibration simulations under four groove damage diameter conditions (0.5334 mm, 0.7112 mm, and 0.5334 mm) are based on vibration data collected at four time stages during the operation of the rolling bearing. (See attached data.) Figure 2-5 .

[0069] Figure 2-5 The vibration data of the rolling bearings at the first to fourth time stages are described. It can be seen that the larger the damage diameter, the more severe the rolling bearing vibration and the worse its service performance. Therefore, the damage and wear of internal components of the rolling bearing can be evaluated by analyzing the vibration performance characteristics.

[0070] The median and average values ​​of the robustness data for the rolling bearings in the first to fourth time stages were calculated based on the mathematical model. The results are shown in [Table / Reference]. Figure 6 .

[0071] Figure 6 The changes in the median and mean of the rolling bearing robustness data from the first to fourth time stages are shown. The trends are consistent, with the median increasing with time, mirroring the trend of the inner raceway damage diameter, thus reflecting the damage status of the inner raceway. Comparing the trends, the median shows a more pronounced trend than the mean. However, the median only reflects the location of the vibration data, providing less information, while the mean contains more data. Based on the above data model, the local true value Ti and service characteristic coefficient Wi of the bearing at each time stage are calculated (see...). Figure 7 , Figure 8 .

[0072] Depend on Figure 7 and Figure 8 As can be seen, when the groove damage diameter is 0mm, the local true value of bearing vibration is 0.05V, and the service factor is only 0, indicating that the rolling bearing is in normal service performance with no signs of performance degradation and almost no possibility of bearing failure. When the groove damage diameter is 0.1778mm, the local true value of bearing vibration is 0.12V, and the service factor is 148, indicating that the rolling bearing is entering an abnormal service performance, its performance is beginning to degrade, and there are signs of bearing failure. The bearing's operating condition should be closely monitored or the bearing should be replaced. When the groove damage diameter is 0.5334mm, the local true value of bearing vibration is 0.17V, and the service factor is 247, indicating that the abnormal state of rolling bearing service performance is gradually intensifying, the performance degradation is gradually worsening, and the potential for bearing failure is increasing. Operation must be stopped and the bearing replaced. When the groove damage diameter is 0.7112mm, the local true value of bearing vibration is 0.322V, and the service factor is 564, indicating that the abnormal state of rolling bearing service performance is rapidly intensifying, the performance degradation is rapidly worsening, the bearing has failed, and a major safety accident may occur.

[0073] Example 2

[0074] To further verify the effectiveness of the above methods and evaluation system, vibration data of the 7008AC / P2 precision rolling bearing (desulfurization pump spindle bearing) were used to evaluate the performance and service status of the precision bearing.

[0075] Test equipment

[0076] Vibration tests on the 7008AC / P2 precision rolling bearing were conducted on the ABLT-1A bearing life strengthening tester. The ABLT-1A bearing life strengthening tester consists of a test spindle, transmission system, loading system, and data acquisition system, and is existing technology, so its structure will not be described in detail.

[0077] The rated dynamic load (Cr) of the precision rolling bearing 700 8AC / P2 is 19kN, and its limiting speed is 12,000 r / min when fully lubricated with oil, with P / C=0.3.

[0078] The axial load (Fa) for the stress test was 4.58 kN, the radial load (Fr) was 4.17 kN, and the rotational speed was 4000 r / min. Vibration data were collected at a rate of 300 Hz, with one data point collected per minute, for a total of 8670 data points.

[0079] Under the above test conditions, vibration data of the 7008AC / P2 precision rolling bearing were collected, as follows: Figure 9 As shown.

[0080] according to Figure 9 During the experiment, the bearing vibration exhibited significant fluctuations, representing a nonlinear dynamic evolution process, falling into the category of information with unknown distribution. As the test time increased, its performance gradually deteriorated, with the trend becoming increasingly pronounced. In the initial stage of bearing operation, between 0 and 3000, the bearing performance was relatively stable. As the operating time increased, between 3000 and 6000, the trend first showed an upward movement followed by a downward fluctuation, indicating that the bearing performance began to change. In subsequent bearing operation, the trend became more significant, demonstrating a clear variation in bearing performance.

[0081] To further analyze the degradation process of the precision rolling bearing 7008AC / P2, data points 0-8000 were used as the research object, and data points 8001-8670 were used for verification. The data points 0-8000 were divided into four time periods, with 2000 data points in each time period (i=4, m=2000). The data were processed according to the method of this invention, and the results are shown in Table 1.

[0082] Table 1 Performance Service Factor of 7008AC / P2 Precision Rolling Bearing

[0083] Serial number i Local True Value of Vibration Ti Service factor Wi 1 9.274 0 2 12.298 32.6 3 12.081 30.3 4 20.132 117

[0084] As shown in Table 1, with the increase of bearing service time, the local true value of bearing vibration and service coefficient increase, indicating that the bearing service performance is deteriorating. The local true value does not change much in the second and third time stages, indicating that the bearing service performance is similar in the second and third time stages. In the fourth time stage, the local true value of bearing vibration and service coefficient increase sharply, and the bearing service performance fails, which is consistent with the experimental results.

[0085] As can be seen from Example 1, by using the method provided by the present invention and based on the experimental data from the bearing data center website of Case Western Reserve University, the vibration local true value and service coefficient constructed by the median and Huber M fusion method can be verified to quantitatively and qualitatively evaluate the service performance of rolling bearings.

[0086] As can be seen from Table 1 of the embodiments, the vibration local true value and service coefficient of the precision rolling bearing 7008AC / P2 increase significantly with the increase of time stages 1 to 4. When it fails, the service coefficient exceeds 100, which is significantly greater than the values ​​of other time stages.

[0087] It is evident that the vibration local true value and service coefficient constructed by the fusion method of median estimation and Huber M estimation can be used to evaluate the service condition of the desulfurization pump spindle bearing, providing an effective method for evaluating the service condition of rolling bearings with unknown distribution characteristics.

[0088] This invention addresses the unknown distribution characteristics of the performance evolution process of desulfurization pump main shaft bearings. It analyzes the robustness of bearing vibration data at different time stages using a median and Huber M fusion method, determines robust bearing vibration data, constructs a bearing performance evaluation system composed of vibration local true values ​​and service coefficients, and establishes an equivalence relationship between the evaluation system and bearing performance standards. This allows for qualitative and quantitative assessment of bearing performance changes.

[0089] The parts of this invention not described in detail are prior art.

Claims

1. A method for evaluating the service performance of a desulfurization pump main shaft bearing, characterized by: The method comprises the following steps: S1, dividing the service life of the desulfurization pump main shaft bearing into m time stages, collecting vibration data of the desulfurization pump main shaft bearing at different time stages during the service, and constituting data sequences of the m time stages; S2, for the vibration data of each time stage, sorting in ascending order to obtain an absolute value sorting sequence; S3, according to statistics, finding the median of each absolute value sorting sequence; S4, according to Huber M estimation principle, obtaining an improved data sequence of each absolute value sorting sequence; S5, according to statistics, obtaining the average value of each improved data sequence; S6, obtaining the absolute difference between the average value of each improved data sequence and the median of the corresponding absolute value sorting sequence; S7, for the vibration data of each time stage, obtaining a local intrinsic interval; S8, according to the local intrinsic interval, calculating the local true value of the vibration of the bearing at each time stage; S9, calculating the service coefficient of the bearing at each time stage; the greater the value of the service coefficient, the worse the service performance.

2. The method for evaluating service performance of a desulfurization pump main shaft bearing according to claim 1, characterized in that: In the step S1, the data sequence of the i-th time stage of the desulfurization pump main shaft bearing during the service is denoted as Xi, and the mathematical expression of Xi is as follows: (1) In formula (1), Xi is a vibration data sequence of a desulfurization pump main shaft bearing in an i th time stage, i is a time stage serial number, n is a data serial number, xi(n) is an n th data in the i th time stage, N is a number of data obtained in the i th time stage, and m is a number of time stages. ​ 3. The method for evaluating service performance of a desulfurization pump main shaft bearing according to claim 1, characterized in that: In the step S2, for the vibration data of the i-th time stage, the absolute value sorting sequence Yi is expressed as follows: (2) In formula (2), Yi is the absolute value sorting sequence, i is the time stage number, n is the data number, N is the number of data obtained at the i-th time stage, m is the number of time stages, and yi(n) is the n-th data in the absolute value sorting sequence.

4. The method for evaluating service performance of a desulfurization pump main shaft bearing according to claim 1, characterized in that: In the step S3, the median of the absolute value sorting sequence of the vibration data of the i-th time stage is specifically expressed as follows: (3) In formula (3), βi is the median of the absolute value sorting sequence, i is the time stage number, N is the number of data obtained at the i-th time stage, m is the number of time stages, yi(n) is the n-th data in the absolute value sorting sequence, n is the data number, n=1, 2, …, N, and N is the number of data obtained at the i-th time stage.

5. The method for evaluating service performance of a desulfurization pump spindle bearing according to claim 1, characterized in that: In the step S4, the improved data sequence of any absolute value sorting sequence is obtained by the following method: Assume that yi(b) and yi(e) are the first and the last data in the ascending order sequence of absolute values, respectively and and are two data in 1, 2, …, N, and yi(b) ≤ βi, βi≤ yi(e); the ascending order sequence of yi(b), …, βi is defined as the left sequence; the number of data in the left sequence is n1; yi(b) is the first data in the left sequence; the ascending order sequence of βi, …, yi(e) is defined as the right sequence; the number of data in the right sequence is n2; yi(e) is the last data in the right sequence;​ According to Huber M estimation principle, when yi(n)≤yi(b), yi(n) is replaced by yi(b); when yi(n)≥yi(e), yi(n) is replaced by yi(e); thus, the improved data sequence is obtained as follows: (4) In formula (4), Zi(n1, n2) is the improved data sequence, zi(n; n1, n2) is the n-th data of the improved data sequence, i is the time stage number, n is the data number, N is the number of data obtained at the i-th time stage, m is the number of time stages, n1 is the number of data of the left sequence, and n2 is the number of data of the right sequence.

6. The method for evaluating service performance of a desulfurization pump spindle bearing according to claim 1, characterized in that: In the step S5, the average value of any improved data sequence is specifically as follows: ； (5) In formula (5), ηi(n1, n2) is the improved data sequence average value, z i(n; n1, n2) is the nth data of the improved data sequence, i is the time stage number, n is the data number, N is the number of data obtained in the ith time stage, m is the number of time stages, n1 is the number of data of the left sequence, and n2 is the number of data of the right sequence.

7. The method for evaluating service performance of a desulfurization pump spindle bearing according to claim 1, characterized in that: In the step S6, the absolute difference between the average value of any improved data sequence and the median of the corresponding absolute value ordering sequence is specifically as follows: ； (6) In formula (6), Di(n1, n2) is the absolute difference between the average value of the improved data sequence and the median of the corresponding absolute value ordering sequence, βi is the median of the absolute value ordering sequence, ηi(n1, n2) is the average value of the improved data sequence, i is the time stage number, m is the number of time stages, n1 is the number of data of the left sequence, and n2 is the number of data of the right sequence.

8. The method for evaluating service performance of a desulfurization pump spindle bearing according to claim 1, characterized in that: In the step S7, for the vibration data of the ith time stage, the specific method for obtaining the local eigenvalue interval [Ki1, Ki2] is as follows: According to the robust characteristics of the median in statistics, when N is even, n1=n2=1, 2, …, N / 2; when N is odd, n1=n2=1, 2, …, (N+1) / 2; N is the number of data obtained in the ith time stage; i is the time stage number; n1 is the number of data of the left sequence; and n2 is the number of data of the right sequence; Different n1 and n2 values are taken to obtain different absolute differences Di(n1, n2) between the average value of the improved data sequence and the median of the absolute value ordering sequence; According to the data robustness theory of modern statistics, for robust data, the significance level is α=(n1+n2) / N, α≤0.1, and the limit value is 0.1; in the range of 0~0.1, in the robust data, the significance level α corresponding to the mean closest to the median is the significance level of the data; therefore, for different Di(n1, n2), the corresponding significance level α is calculated, the Di(n1, n2) with the smallest significance level α is found, which is expressed as Dimin, the first data yi(b) of the left sequence and the tail data yi(e) of the right sequence corresponding to Dimin are Ki1 and Ki2 respectively; the local eigenvalue interval of the ith time stage is [Ki1, Ki2]; Ki1 is the lower limit value of the local eigenvalue interval; and Ki2 is the upper limit value of the local eigenvalue interval.

9. The method for evaluating service performance of a desulfurization pump spindle bearing according to claim 1, characterized in that: In the step S8, the mathematical expression of the vibration local true value Ti of the bearing in the ith time stage is specifically as follows: (7) 10. The method for evaluating service performance of a desulfurization pump spindle bearing according to claim 1, characterized in that: In the step S9, the mathematical expression of the service coefficient of the bearing in the ith time stage is specifically as follows: (8) In formula (8): is the minimum value of Ti, Ti is the local true value of vibration in the i th time stage; Wi is the service coefficient of the bearing in the i th time stage, and the larger the value of Wi is, the worse the service performance of the bearing is.

Images