Thrust bearing full-life fault diagnosis method and system based on film thickness and wear

Abstract

The present application relates to a thrust bearing full-life fault diagnosis method and system based on film thickness and wear. The method comprises: first establishing failure probability models for two stages, i.e., a lubrication failure stage and a wear failure stage, and simultaneously representing a lubrication state and a wear state of a bearing in the form of probabilities; then, performing a normal distribution test on monitoring data, and performing normal distribution conversion on abnormal distribution data; and finally, by using the established failure probability models for the two stages, i.e., the lubrication failure stage and the wear failure stage, and the acquired normal distribution data, performing diagnosis on the lubrication state and the wear state of the bearing on the basis of a 3σ criterion, thereby realizing full-life fault monitoring. In the present invention, failure evolution situations of two stages, i.e., a lubrication stage and a wear stage, of a thrust bearing are represented in the form of probabilities, so as to realize the full-life fault diagnosis of the thrust bearing, thereby addressing the current gap in research on dynamic fault diagnosis throughout the full-life cycle of thrust bearings, and improving the accuracy of fault early warning and condition-based maintenance of the thrust bearings under actual working conditions.

Description

Thrust bearing full life fault diagnosis method and system based on film thickness and wear

[0001] Cross-references

[0002] This application refers to Chinese patent application No. 2024105689090, filed on May 9, 2024, entitled “Thrust bearing full-life fault diagnosis method and system based on film thickness and wear”, which is incorporated into this application in its entirety by reference. Technical Field

[0003] The present application relates to the technical field of thrust bearing fault detection, and in particular to a method and system for diagnosing full-life faults of thrust bearings based on film thickness and wear. Background Art

[0004] Thrust bearings play a vital role in vertical equipment such as nuclear main pumps and hydro-turbine units, supporting the axial load of the entire unit and being a key component for ensuring safe, reliable, and long-term stable operation. Thrust bearing failures primarily occur in two stages: early lubrication failure and later bearing wear failure. The first stage is characterized by a shift in lubrication state from hydrodynamic lubrication to mixed lubrication due to oil film thinning. This failure can induce bearing rubbing, bearing wear, and bearing burnout. The second stage is characterized by localized bearing wear due to poor lubrication. This irreversible wear can accumulate and lead to serious failures such as the removal and shedding of babbitt metal from the bearing surface. Therefore, monitoring the lubrication and wear status of thrust bearings is of great engineering significance for fault warning and condition-based maintenance throughout the unit's lifecycle.

[0005] For lubrication state monitoring, vibration, acoustic emission and other monitoring methods are currently used; however, vibration monitoring and acoustic emission monitoring methods ([1]Xing PF, Li GB, Gao H, & Wang, G.Experimental investigation on identifying friction state in lubricated tribosystem based on friction-induced vibration signals[J].Mechanical systems and signal processing, 2020, 138:106590[2]Mokhtari N, Pelham JG, Nowoisky S.Friction and wear monitoring methods for journal bearings of geared turbofans based on acoustic emission signals and machine learning[J].Lubricants, 2020, 8(3):29.) mainly use the signals generated by friction contact to reflect the poor lubrication phenomenon, which cannot cover the evolution process of the lubrication state before contact. The oil film thickness monitoring method mainly uses the oil film thickness measurement value to monitor the bearing lubrication status in real time, which can realize early failure detection ([3] Dou P, Jia Y, Zheng P, Wu, T, Yu, M, Reddyhoff, T, & Peng, Z. Review of ultrasonic-based technology for oil film thickness measurement in lubrication[J]. Tribology International, 2022, 165: 107-290.). In terms of wear measurement, there are mainly two types of methods: online and offline. Offline measurement mainly reflects the degree of wear by measuring the loss of mass before and after wear or measuring the profile of the surface after wear, but it requires the equipment to be shut down and the diagnostic results have a lag. Online measurement can use eddy current sensors, linear potentiometers, laser displacement sensors, etc. to reflect the degree of wear by measuring the changes in component position and displacement before and after wear, but these sensors will damage the structure of the bearing when installed.

[0006] Ultrasonic technology, with its non-invasive characteristics, can achieve online measurement of lubricating film thickness without disturbing the lubrication state or damaging the bearing structure. At the same time, in terms of wear measurement, ultrasonic sensors have been successfully used to achieve real-time measurement of the wear depth of the bearing lining on thrust bearings: the thickness of the bearing lining after wear is measured in real time using the time-of-flight method, and the wear depth of the bearing lining is obtained by subtracting the thickness after wear from the original thickness of the bearing lining ([4] Brunskill H, Harper P, Lewis R. The real-time measurement of wear using ultrasonic reflectometry [J]. Wear, 2015, 332: 1129-1133.).

[0007] However, existing ultrasonic technologies monitor only a single variable, wear or film thickness, and determine the bearing's current condition by comparing it against theoretical or pre-set thresholds. This allows monitoring only during the lubrication or wear failure phase, making it difficult to diagnose faults throughout the entire operating lifecycle. Furthermore, the high sampling frequency of lubrication oil film monitoring data results in a large degree of data dispersion, and wear failure presents different failure risk probabilities at different times during the degradation process. Therefore, simply using a given threshold for diagnosis is less robust, and timely and accurate diagnosis is difficult to ensure. Therefore, achieving dynamic and accurate fault diagnosis throughout the entire lifecycle of thrust bearings remains a major challenge.

[0008] Summary of the Invention

[0009] Based on this, the present application provides a thrust bearing full-life fault diagnosis method and system based on film thickness and wear, which characterizes the failure evolution of the thrust bearing in the two stages of lubrication and wear in the form of probability, realizes fault diagnosis of the thrust bearing throughout its life cycle, solves the current gap in research on dynamic fault diagnosis of thrust bearings throughout their life cycle, and improves the accuracy of thrust bearing fault warning and situation-based maintenance under actual working conditions.

[0010] In a first aspect, the present application provides a method for diagnosing full-life thrust bearing faults based on film thickness and wear. The technical solution of the embodiment of the present application includes the following steps:

[0011] S1. Establish a two-stage failure probability model of lubrication and wear failure, and characterize the lubrication and wear states of the bearing in a probabilistic manner.

[0012] S2. Perform a normal distribution test on the monitoring data. If the data conforms to the normal distribution, proceed directly to step S3; for non-normally distributed data, perform a normal distribution conversion and then proceed to step S3;

[0013] S3. Using the two-stage failure probability model of lubrication and wear failure established in step S1 and the normal distribution data obtained in step S2, the lubrication state and wear state of the bearing are diagnosed based on the 3σ criterion to achieve full-life fault monitoring.

[0014] In some embodiments, the lubrication failure probability model established in step S1 is specifically:

[0015] Where x1 is the current monitoring value of the film thickness, Ф1(x1) is the probability of lubrication failure, and f1(x1) is the probability density function of lubrication failure; D1 is the theoretical threshold of lubrication failure, which is equal to the surface roughness of the bearing in the lubrication failure stage; σ1 is the standard deviation of the film thickness, which is obtained by calculating the standard deviation of the complete data obtained in the experiment.

[0016] In some embodiments, the wear failure probability model established in step S1 is specifically:

[0017] Where x2 is the current monitoring value of the wear depth, Φ2(x2) is the wear failure probability, and f2(x2) is the wear failure probability density function; D2 is the theoretical threshold of wear failure, which is equal to the thickness of the bearing lining in the wear failure stage; σ2 is the standard deviation of the wear depth, which is obtained by calculating the standard deviation of the complete data obtained in the experiment.

[0018] In some embodiments, in step S2, the normality hypothesis is tested using the Jarque-Bera test method. If the normal distribution is not met, a Box-Cox transformation is used to transform it into a state close to the normal distribution. The Jarque-Bera test method formula is specifically:

[0019] Among them, n is the sample size of indicator monitoring data, S is the skewness of the sample, and K is the kurtosis of the sample.

[0020] In some embodiments, the skewness and kurtosis of the monitoring indicator data set are first calculated, and then the Jarque-Bera statistic is calculated using the formula, and the JB statistic is compared with the critical value of the chi-square distribution to test whether the monitoring data conforms to the normal distribution.

[0021] In some embodiments, the specific method of Box-Cox transformation is:

[0022] Where x is the original monitoring data, which must be a positive number, y(λ) is the transformed data, and λ is the transformation parameter obtained through data optimization, with the aim of making the transformed data as close to the normal distribution as possible.

[0023] In some embodiments, the 3σ principle is introduced in step S3 to formulate a bearing diagnosis strategy: that is, when the lubrication and wear failure probabilities Φ(x) are both less than 0.6826, the diagnostic bearing is in a good operating state; when one of the two Φ(x) is greater than or equal to 0.6826, the diagnostic bearing is in a warning state, and the diagnostic result lubrication / wear failure warning is output; when Φ(x) in one of the two stages is greater than 0.9545, it means that lubrication failure or wear failure has occurred.

[0024] In a second aspect, the present application provides a full-life fault diagnosis system for thrust bearings based on film thickness and wear, including: a model module: constructing a two-stage failure probability model of lubrication and wear failure, and simultaneously characterizing the lubrication state and wear state of the bearing in the form of probability;

[0025] Data verification module: performs normal distribution test on monitoring data and performs normal distribution conversion on non-normal distribution data obtained;

[0026] Diagnosis module: Using the failure probability model established by the model module and the normal or nearly normal distribution data obtained by the data verification module, the lubrication and wear conditions of the bearing are diagnosed based on the 3σ criterion to achieve full-life fault diagnosis.

[0027] Compared with the prior art, the present invention has at least the following beneficial effects:

[0028] The method of the present invention first establishes a two-stage failure probability model for lubrication and wear failure, simultaneously characterizing the lubrication and wear states of the bearing in a probabilistic manner. The monitoring data is then tested for normal distribution, and any non-normally distributed data obtained is converted to a normal distribution. Finally, the established model and data are used to diagnose the lubrication and wear states of the bearing based on the 3σ criterion. Because the full-life degradation process of a thrust bearing includes two stages, lubrication and wear failure, the first stage is characterized by lubrication state migration caused by oil film thinning, and lubrication state degradation can be diagnosed by measuring the oil film thickness. The second stage is characterized by localized wear of the bearing shell due to poor lubrication. This irreversible wear accumulates and leads to the removal of babbitt metal from the bearing shell surface. The wear depth of the bearing shell lining allows for fault diagnosis during the wear stage. Therefore, the method of the present invention can comprehensively implement fault warnings for the entire life cycle of a thrust bearing, which is of great engineering significance.

[0029] The present invention targets the two stages of lubrication failure and wear failure, and the established probability models of lubrication failure and wear failure can characterize the monitoring indicators in a probabilistic manner, and use the data obtained from the experiment to construct data-driven and optimized thresholds and model parameters, which can improve the fit of the model and the robustness of the diagnosis.

[0030] In order to make the data meet the model assumptions, the present invention uses the Jarque-Bera test method to perform a normal distribution hypothesis test on the data. If it does not meet the normal distribution, the Box-Cox transformation is used to convert it into a state close to the normal distribution. Since the sampling data set is large, if the conventional hypothesis testing method is adopted, the deviation of individual data from the normal distribution may cause the test to fail, and these data are sometimes unimportant for the entire system, which will reduce the scientific nature of the hypothesis test results. The Jarque-Bera test method is to compare the skewness and kurtosis of the sample data with the skewness and kurtosis of the normal distribution (the skewness of the normal distribution is 0 and the kurtosis is 3), which is more accurate for the test of large data sets.

[0031] This invention incorporates the 3σ principle to develop a scientific bearing diagnosis strategy: when both the lubrication and wear failure probabilities Φ(x) are less than 0.6826, the bearing is diagnosed as being in good operating condition; when either Φ(x) is greater than or equal to 0.6826, the bearing is in a warning state, and a lubrication / wear failure warning is output. When Φ(x) in either stage is greater than 0.9545, lubrication failure or wear failure has occurred. In a normal distribution, approximately 68.26% of the data lies within one standard deviation of the mean, and 95.45% of the data lies within two standard deviations. This method, based on the data distribution characteristics, provides a scientific basis for bearing maintenance. Furthermore, by setting failure probability thresholds, it is possible to effectively distinguish when maintenance should be performed and when warnings should be issued, thus avoiding excessive maintenance, saving maintenance costs, and ensuring reliable bearing operation. Furthermore, using the two-stage failure probability thresholds, lubrication failure can be distinguished from wear failure, helping to determine more targeted maintenance measures and improving maintenance effectiveness.

[0032] In summary, the method of the present invention realizes fault diagnosis of the entire life cycle of the thrust bearing, solves the current blank in the research of dynamic fault diagnosis of the entire life cycle of the thrust bearing, and improves the accuracy of thrust bearing fault warning and situation-based maintenance under actual working conditions.

[0033] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the embodiments below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present application. The same reference numerals are used throughout the drawings to denote the same components. In the drawings:

[0035] FIG1 is a flow chart of a method for diagnosing full-life faults of a thrust bearing based on film thickness and wear according to an embodiment of the present application;

[0036] FIG2 is a diagram of diagnostic results provided in another embodiment of the present application. DETAILED DESCRIPTION

[0037] The following embodiments of the technical solution of the present application will be described in detail with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present application and are therefore only examples and are not intended to limit the scope of protection of the present application.

[0038] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned figure descriptions are intended to cover non-exclusive inclusions.

[0039] 1 , a method for diagnosing full-life faults of a thrust bearing based on film thickness and wear includes the following steps:

[0040] S1. Establish a two-stage failure probability model of lubrication and wear failure, and characterize the lubrication and wear states of the bearing in a probabilistic manner.

[0041] Assuming that the change in oil film thickness and the degradation of bearing liner wear follow a normal distribution, a unified failure probability model for the two stages is established. The specific lubrication failure probability model established is:

[0042] Where x1 is the current monitoring value of the film thickness, Φ1(x1) is the probability of lubrication failure, and f1(x1) is the probability density function of lubrication failure; D1 is the theoretical threshold of lubrication failure, which is equal to the surface roughness of the bearing in the lubrication failure stage; σ1 is the standard deviation of the film thickness, which is obtained by calculating the standard deviation of the complete data obtained in the experiment.

[0043] The established wear failure probability model is as follows:

[0044] Where x2 is the current monitoring value of the wear depth, Φ2(x2) is the wear failure probability, and f2(x2) is the wear failure probability density function; D2 is the theoretical threshold of wear failure, which is equal to the thickness of the bearing lining in the wear failure stage; σ2 is the standard deviation of the wear depth, which is obtained by calculating the standard deviation of the complete data obtained in the experiment.

[0045] S2. Perform a normal distribution test on the monitoring data. If the data conforms to the normal distribution, proceed directly to step S3; for non-normally distributed data, perform a normal distribution conversion and then proceed to step S3;

[0046] Due to the large size of the monitoring data set, the Jarque-Bera test method was used for the normal hypothesis. When the hypothesis test passed, the two-stage failure probability formula was used to characterize the bearing lubrication and wear status. When the hypothesis test failed, the monitoring data set was transformed by Box-Cox to make the data characteristics close to the normal distribution. The failure probability was then calculated based on the two-stage failure probability formula.

[0047] The Jarque-Bera test method formula is as follows:

[0048] Where n is the sample size of the indicator monitoring data, S is the skewness of the sample, and K is the kurtosis of the sample;

[0049] First, the skewness and kurtosis of the monitoring indicator data set are calculated. Then, the Jarque-Bera statistic is calculated using the formula. The JB statistic is compared with the critical value of the chi-square distribution to test whether the monitoring data conforms to the normal distribution.

[0050] When the hypothesis test fails, the experimental data set is transformed into a Box-Cox transformation to make the data characteristics close to the normal distribution. The specific method of Box-Cox transformation is as follows:

[0051] Among them, x is the original monitoring data (must be a positive number), y(λ) is the transformed data, and λ is the transformation parameter, which can be obtained through data optimization. The purpose is to make the transformed data as close to the normal distribution as possible.

[0052] S3. Using the two-stage failure probability model of lubrication and wear failure established in step S1 and the normal distribution data obtained in step S2, the lubrication state and wear state of the bearing are diagnosed based on the 3σ criterion to achieve full-life fault monitoring.

[0053] The 3σ principle is introduced to formulate a scientific bearing diagnosis strategy: that is, when the probability of lubrication and wear failure Φ(x) are both less than 0.6826, the diagnostic bearing is in good operating condition; when one of the two Φ(x) is greater than or equal to 0.6826, the diagnostic bearing is in a warning state and the diagnostic result lubrication / wear failure warning is output; when one of the two stages Φ(x) is greater than 0.9545, it means that lubrication failure or wear failure has occurred.

[0054] Referring to Figure 2, a diagnostic case of this embodiment is shown, in which the boxes represent the actual bearing status and the dots represent the diagnosed bearing status. It can be seen from Figure 2 that the method of this embodiment can capture all bearing failure states in the embodiment, and the false detection rate is low in a good state.

[0055] A full-life fault diagnosis system for thrust bearings based on film thickness and wear, comprising:

[0056] Model module: Constructs a two-stage failure probability model of lubrication and wear failure, and simultaneously characterizes the lubrication state and wear state of the bearing in the form of probability;

[0057] Data verification module: performs normal distribution test on monitoring data and performs normal distribution conversion on non-normal distribution data obtained;

[0058] Diagnosis module: Using the failure probability model established by the model module and the normal or nearly normal distribution data obtained by the data verification module, the lubrication and wear conditions of the bearing are diagnosed based on the 3σ criterion to achieve full-life fault diagnosis.

[0059] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Furthermore, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0060] The present application is described with reference to the flow chart and / or block diagram of the method, device (system), and computer program product according to the embodiment of the present application. It should be understood that each flow process and / or box in the flow chart and / or block diagram and the combination of the flow process and / or box in the flow chart and / or block diagram can be realized by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processing machine or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for realizing the function specified in one flow chart flow or multiple flows and / or one box or multiple boxes of the block diagram.

[0061] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction device that implements the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0062] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0063] The above content is only for explaining the technical idea of ​​the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and they should all be included in the scope of the claims and specification of the present application. In particular, as long as there is no structural conflict, the various technical features mentioned in the various embodiments can be combined in any way. The present application is not limited to the specific embodiments disclosed herein, but includes all technical solutions that fall within the scope of the claims.

Claims

1. A method for full-life fault diagnosis of thrust bearings based on film thickness and wear, wherein: The following steps are involved: S1. Establish a two-stage failure probability model of lubrication and wear failure, and characterize the lubrication state and wear state of the bearing in the form of probability; S2, perform a normal distribution test on the monitoring data. If the data conforms to the normal distribution, proceed directly to step S3; for non-normal distribution data, perform a normal distribution conversion, and then proceed to step S3; S3. Using the two-stage failure probability model of lubrication and wear failure established in step S1 and the normal distribution data obtained in step S2, the lubrication state and wear state of the bearing are diagnosed based on the 3σ criterion to achieve full-life fault monitoring.

2. The method according to claim 1, wherein: The lubrication failure probability model established in step S1 is specifically: Where x1 is the current monitoring value of the film thickness, Φ1(x1) is the probability of lubrication failure, f1(x1) is the probability density function of lubrication failure; D1 is the theoretical threshold of lubrication failure, which is equal to the surface roughness of the bearing in the lubrication failure stage; σ1 is the standard deviation of the film thickness, which is obtained by calculating the standard deviation of the complete data obtained in the experiment.

3. The method according to claim 1, wherein: The wear failure probability model established in step S1 is specifically: Where x2 is the current monitoring value of the wear depth, Φ2(x2) is the probability of lubrication failure, f2(x2) is the probability density function of lubrication failure; D2 is the theoretical threshold of wear failure, which is equal to the thickness of the bearing lining in the wear failure stage; σ2 is the standard deviation of the wear depth, which is obtained by calculating the standard deviation of the complete data obtained in the experiment.

4. The method according to claim 1, wherein: In step S2, the normal hypothesis adopts the Jarque-Bera test method. If it does not conform to the normal distribution, the Box-Cox transformation is used to transform it into a state close to the normal distribution. The specific formula of the Jarque-Bera test method is: Among them, n is the sample size of indicator monitoring data, S is the skewness of the sample, and K is the kurtosis of the sample.

5. The method according to claim 4, wherein: First, the skewness and kurtosis of the monitoring indicator data set were calculated, and then the Jarque-Bera statistic was calculated using the formula. The JB statistic was compared with the critical value of the chi-square distribution to test whether the monitoring data conformed to the normal distribution.

6. The method according to claim 4, wherein: The specific method of Box-Cox transformation is: Where x is the original monitoring data, which must be a positive number, y(λ) is the converted data, and λ is the conversion parameter obtained through data optimization, the purpose of which is to make the converted data as close to the normal distribution as possible.

7. The method according to claim 1, wherein: In step S3, the 3σ principle is introduced to formulate the bearing diagnosis strategy: that is, when the probability of lubrication and wear failure Φ(x) are both less than 0.6826, the bearing is diagnosed to be in good operating condition; when one of the two Φ(x) is greater than or equal to 0.6826, the bearing is diagnosed to be in a warning state, and the diagnosis result lubrication / wear failure warning is output; when one of the two stages Φ(x) is greater than 0.9545, it means that lubrication failure or wear failure has occurred.

8. A thrust bearing full-life fault diagnosis system based on film thickness and wear that implements the method described in any one of claims 1 to 7, wherein: include: Model module: Construct a two-stage failure probability model of lubrication and wear failure, and characterize the lubrication state and wear state of the bearing in the form of probability; Data verification module: perform normal distribution test on monitoring data, and perform normal distribution conversion on non-normal distribution data obtained; Diagnosis module: Using the failure probability model established by the model module and the normal or nearly normal distribution data obtained by the data verification module, the lubrication and wear conditions of the bearing are diagnosed based on the 3σ criterion to achieve full-life fault diagnosis.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, wherein: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images