Method and system for predicting online residual life of double-variable wet clutch

Abstract

The invention discloses a bivariate wet clutch online residual life prediction method and system, and the method comprises the steps: obtaining the real-time friction torque data of a clutch, carrying out the preprocessing of the real-time friction torque data, and extracting key performance characteristic parameters; a bivariate inverse Gaussian degradation model is constructed based on the key performance characteristic parameters, dynamic online prediction of the clutch RUL is carried out through the bivariate inverse Gaussian degradation model, parameter estimation is carried out on the bivariate inverse Gaussian degradation model through a Markov chain Monte Carlo MCMC method, and dynamic online prediction of the clutch RUL is carried out. On-line prediction of the clutch RUL is realized based on mass sampling of a Monte Carlo simulation method. According to the method, flexible fitting of the degradation process and dynamic and accurate online prediction of the remaining service life are achieved, accurate model support is provided for decision making according to conditions, the health management and maintenance plan of the system is optimized, and sudden failures are reduced.

Description

Technical Field

[0001] The invention relates to the interdisciplinary technical field combining equipment reliability engineering with random processes, and in particular to an online remaining life prediction method and system for a dual-variable wet clutch. Background Art

[0002] The wet clutch is a widely used power transmission device, commonly used in automatic transmissions, heavy machinery, high-performance vehicles and other equipment. With the advancement of technology and the increase in demand for use, the working quality and reliability of the wet clutch are increasing. The wet clutch achieves power transmission and switching by controlling the engagement or separation between the friction plate and the steel plate, and often faces high-load and long-term operation scenarios. Therefore, under high load and frequent use, the accumulated friction and thermal effects cause the friction elements of the clutch to gradually wear out, and its torque transmission capacity decreases, which seriously affects the normal working quality and reliability of the clutch and the service life and safety of the vehicle transmission system. Therefore, effective monitoring of the degradation data of the wet clutch, namely the performance characteristic (PC), can not only effectively reveal the degradation law of the wet clutch, but also provide a reliable historical information basis for effective remaining useful life (RUL) prediction. By establishing an accurate prediction model, the service life and reliability of the wet clutch can be extended and the maintenance cost during the use phase can be reduced.

[0003] Degradation analysis, including degradation modeling and parameter estimation, has proven to be an important toolkit, especially for equipment with limited testing time and sample size. A suitable degradation model is crucial for clutch degradation characteristics and reliability representation. In existing reliability analysis methods, univariate models are usually used to describe the degradation process of the clutch system. However, the degradation of the clutch in actual work has multivariate characteristics, and there may be correlations between the variables. In addition, most of the existing bivariate or multivariate models use deterministic models or models based on simple random processes, which have certain limitations in capturing the uncertainty of degradation and its dynamic behavior. In recent years, the inverse Gaussian (IG) process model has attracted widespread attention due to its ability to effectively describe the stochasticity of the system degradation process. However, when dealing with multivariate degradation processes and real-time RUL prediction of complex systems, the existing methods cannot fully cope with the changes in real-time dynamic data, resulting in inaccurate predictions. Especially for online RUL prediction, the traditional estimation methods have low accuracy, resulting in the inability to give accurate RUL for effective preventive maintenance arrangements, increasing sudden failures and maintenance costs. Existing methods lag behind in reliability assessment and preventive maintenance decision support, making it difficult to make effective preventive maintenance arrangements through accurate life prediction, resulting in reduced system reliability and increased maintenance costs. A model that can provide accurate support for decision-making is needed to optimize the system's health management and maintenance plan and reduce sudden failures. Summary of the invention

[0004] To solve the above technical problems, the present invention proposes a dual-variable wet clutch online remaining service life prediction method and system, constructs a dual-variable inverse Gaussian process degradation model based on the real-time online characteristic parameter data of the clutch, and considers the correlation and uncertainty of the degradation variables at the same time, to achieve flexible fitting of the degradation process and dynamic and accurate online prediction of the remaining service life, provide precise model support for situation-based decision-making, optimize the system's health management and maintenance plan, and reduce sudden failures.

[0005] On the one hand, to achieve the above object, the present invention provides a dual variable wet clutch online remaining life prediction method, comprising:

[0006] Acquiring clutch real-time friction torque data, and preprocessing the real-time friction torque data to extract key performance characteristic parameters;

[0007] A two-variable inverse Gaussian degradation model is constructed based on the key performance characteristic parameters, and the dynamic online prediction of the clutch RUL is performed through the two-variable inverse Gaussian degradation model, wherein the parameters of the two-variable inverse Gaussian degradation model are estimated through the Markov chain Monte Carlo MCMC method, and the online prediction of the clutch RUL is achieved based on a large number of samplings of the Monte Carlo simulation method.

[0008] Preferably, obtaining the real-time friction torque data of the clutch and preprocessing the real-time friction torque data include:

[0009] The friction torque data is collected in real time by a speed torque sensor, and the friction torque data is denoised and smoothed to eliminate noise and abnormal points.

[0010] Preferably, the method for extracting the key performance characteristic parameters is:

[0011]

[0012] t e =t b -t a ;

[0013] Where, t e is the engagement time; s is the number of friction pairs; p is the applied pressure; R o and R i are the outer radius and inner radius of the friction part respectively; t a and t d are the start and end times of the engagement process, T(t) is the real-time friction torque, and μ is the friction coefficient COF.

[0014] Preferably, constructing the bivariate inverse Gaussian degradation model includes:

[0015] Obtaining performance characteristic parameters through friction torque data, and assuming that random variables are controlled by probability distribution, thereby obtaining a probability density function of the degradation process, wherein the performance characteristic parameters include the degradation process of friction coefficient COF and engagement time;

[0016] By setting n clutch units to perform degradation tests under the same conditions, each unit i obtains m degradation observations. Given the latent variables of the degraded unit, the posterior distribution of the latent variables is obtained, and finally the complete log-likelihood function of the bivariate inverse Gaussian degradation model is obtained.

[0017] Preferably, the probability density function of the degradation process is:

[0018] The extracted friction coefficient COF and engagement time are set to be Y1(t) and Y2(t), t≥0, and they obey the inverse Gaussian distribution, that is, Y p (t)~IG(μΛ p (t), η p Λ p (t) 2 ), where p = 1, 2, representing two degradation processes respectively; μ is a common latent variable used to describe the dependency between the two degradation processes; Λ(t) is a time-dependent shape function in the form of a power law function. t, q are time and the power of time respectively; η describes the volatility of degradation, thereby obtaining the probability density function of the degradation process:

[0019]

[0020] In the formula, f(Y p ) is the probability density function of the corresponding performance characteristic parameters.

[0021] Preferably, obtaining the probability density function of the degradation process includes:

[0022] Assume that the latent variable μ -1 Obey the truncated normal distribution TN(ω,κ -2 ), and is statistically independent of the parameter η; then assuming that μ -1 , the two performance parameter degradation processes Y1(t) and Y2(t) are independent of each other, so the degradation process can be expressed as Y p (t)|μ -1 ~IG(μΛ p (t), η p Λ p (t) 2 ), p = 1, 2, so the probability density function of the degradation process is:

[0023] f(Y k (t)|μ -1 )=f(Y p );

[0024] The probability density function of the latent variable is:

[0025]

[0026] In the formula, π(μ -1 ) is the probability density function of the latent variable, ω is the mean parameter, κ is the standard deviation parameter, φ is the probability density function of the standard normal distribution, and Φ is the cumulative distribution function of the standard normal distribution.

[0027] Preferably, the complete log-likelihood function of the bivariate inverse Gaussian degradation model is:

[0028]

[0029] Where θ = (ω, κ, η1, η2, q1, q2) is the parameter set of the model to be estimated, and P(·) is the latent variable of a given degradation unit i. The likelihood function of the model, is the m observation data of the friction coefficient COF of unit i, is the m observation data of the joining time of unit i, is the common latent variable of the i-th clutch unit, and N is the number of clutch units.

[0030] Preferably, the dynamic online prediction of the clutch RUL is performed by using the dual variable inverse Gaussian degradation model, including:

[0031] Based on the Monte Carlo simulation method, the degradation data time window of the obtained performance parameters is used to realize the iterative prediction of the RUL of the clutch system, specifically:

[0032] S1. Input known historical data at the current time t into the bivariate inverse Gaussian degradation model, and obtain the posterior distribution range of each parameter based on the MCMC method, including the mean and variance;

[0033] S2. When the latest data does not exceed the preset failure threshold of the key performance characteristic parameter, the most recent historical performance characteristic parameter data is selected to form a degradation data time window, and a Monte Carlo simulation method is used to generate a simulated degradation path of the key performance characteristic parameter based on the model parameter value;

[0034] S3, sampling the parameter interval in S1 S times to obtain S simulated degradation paths, and then calculating the failure time of the simulated degradation paths respectively;

[0035] S4. Obtain the remaining service life of the current clutch unit according to the estimated mean failure time and the current time.

[0036] Preferably, the method further includes comparing the prediction performance of the two-variable inverse Gaussian degradation model with that of the two-variable Gamma model, respectively calculating the MAE of the two models, performing quantitative analysis on the prediction performance, and verifying the effectiveness of the two-variable inverse Gaussian degradation model.

[0037] On the other hand, to achieve the above-mentioned purpose, the present invention also provides a dual variable wet clutch online remaining life prediction system, which is applied to the dual variable wet clutch online remaining life prediction method, comprising:

[0038] Sample torque storage module: used to obtain real-time friction torque data of the clutch, obtain friction torque curve, and store it in the torque storage module;

[0039] Torque preprocessing module: used to denoise and smooth the collected friction torque data, eliminate noise and abnormal points, and extract key performance characteristic parameters based on the smoothed torque data;

[0040] Bivariate Inverse Gaussian Model Module: Constructs a bivariate inverse Gaussian process model based on the processed data;

[0041] Goodness of fit test module: used to evaluate the parameter and path fitting effect estimated by the bivariate inverse Gaussian process model;

[0042] RUL Online Prediction Module: Based on the updated model and real-time data, the Monte Carlo simulation method is used to sample and predict the future trend of the degradation path, and the remaining service life of the wet clutch is dynamically calculated.

[0043] Compared with the prior art, the present invention has the following advantages and technical effects:

[0044] The method of the present invention extracts main performance parameters (friction coefficient and engagement time) according to the real-time friction torque data during clutch operation, and accurately captures the correlation between friction plate wear and friction coefficient change in the wet clutch by constructing a two-variable inverse Gaussian (IG) degradation model. In addition, the method combines Markov chain Monte Carlo (MCMC) and Monte Carlo simulation techniques to achieve high-precision parameter estimation and real-time remaining useful life (RUL) prediction.

[0045] The system proposed in the present invention not only improves the dynamic response capability of life prediction, but also can guide the formulation of situation-based maintenance decisions, reduce the risk of sudden failures, and improve the operational reliability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0047] Figure 1 A flow chart of a dual variable wet clutch online remaining life prediction method according to an embodiment of the present invention;

[0048] Figure 2 It is a structural schematic diagram of a dual variable wet clutch online remaining life prediction system according to an embodiment of the present invention;

[0049] Figure 3 This is a smoothed friction torque diagram of an embodiment of the present invention;

[0050] Figure 4 Graph showing the evolution of COF and bonding time according to an embodiment of the present invention;

[0051] Figure 5 Schematic diagram of RUL online prediction results according to an embodiment of the present invention. DETAILED DESCRIPTION

[0052] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0053] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0054] The working mode of the clutch is mainly to achieve power transmission or interruption by controlling the engagement and separation between the friction plate and the dual steel plate. When the clutch is engaged, the pressure applied by the hydraulic system causes the piston to push the friction plate to contact the dual steel plate, and the power is transmitted through friction; when the clutch is disengaged, the piston returns after the pressure of the hydraulic system is released, the friction plate and the steel plate are separated, and the power transmission is interrupted. However, during the operation of the clutch, the accumulated wear of the friction elements will gradually affect its performance. This dynamic change makes the degradation of the clutch uncertain. In order to effectively evaluate the state and life of the clutch, it becomes very necessary to extract and analyze the key characteristic parameters. During the operation of the clutch, the coefficient of friction (COF) determines the clutch's ability to transmit torque, and the engagement time reflects the clutch's response speed and engagement smoothness. These parameters can directly reflect the working condition of the clutch and help to establish an accurate degradation model, thereby providing data support for real-time remaining life prediction and preventive maintenance.

[0055] The traditional wet clutch dual-variable degradation model and online remaining life prediction method have the following problems, including: 1. Modeling problem of multivariable degradation process: The performance degradation of wet clutch is usually characterized by multiple parameters, and there is a certain correlation between these variables. Existing single variable or simple multivariable models cannot accurately describe the multidimensional degradation behavior of wet clutch. A dual-variable model that can effectively capture multiple degradation variables and their correlations is needed; 2. Insufficient parameter estimation accuracy: In the wet clutch degradation modeling process, the existing methods have the problem of low accuracy when dealing with parameter estimation of complex degradation variables, and cannot fully reflect the degradation characteristics under actual working conditions. It is necessary to introduce efficient numerical methods such as MCMC and Monte Carlo simulation to improve the accuracy of dual-variable model parameter estimation; 3. Real-time remaining life prediction problem: Traditional life prediction methods are usually based on a large amount of historical data and life data, and cannot fully cope with the changes in real-time dynamic data, resulting in inaccurate prediction. A model that can dynamically adjust using real-time data is needed to achieve accurate remaining life (RUL) prediction of wet clutches under actual working conditions; 4. System reliability and maintenance optimization issues: Existing methods lag behind in reliability assessment and preventive maintenance decision support, making it difficult to make effective preventive maintenance arrangements through accurate life prediction, resulting in reduced system reliability and increased maintenance costs. A model that can provide accurate support for decision-making is needed to optimize the system's health management and maintenance plan and reduce sudden failures.

[0056] Based on this, this embodiment proposes a dual variable wet clutch online remaining life prediction method, which has solved many of the above problems, such as Figure 1 ,include:

[0057] Acquiring clutch real-time friction torque data, and preprocessing the real-time friction torque data to extract key performance characteristic parameters;

[0058] A two-variable inverse Gaussian degradation model is constructed based on the key performance characteristic parameters, and the dynamic online prediction of the clutch RUL is performed through the two-variable inverse Gaussian degradation model, wherein the parameters of the two-variable inverse Gaussian degradation model are estimated through the Markov chain Monte Carlo MCMC method, and the online prediction of the clutch RUL is achieved based on a large number of samplings of the Monte Carlo simulation method.

[0059] This embodiment establishes a dual variable (including friction coefficient and engagement time) degradation process model that can capture the performance degradation of the clutch, combines the clutch characteristic parameters collected online in real time, and uses numerical methods such as Markov Chain Monte Carlo (MCMC) and Monte Carlo simulation to achieve accurate online RUL prediction of the clutch system during its service.

[0060] Further, obtaining the real-time friction torque data of the clutch and preprocessing the real-time friction torque data include:

[0061] The real-time friction torque data T(t) is collected through the speed torque sensor in the clutch system; the collected friction torque data of the engagement process is denoised and smoothed to eliminate noise and abnormal points.

[0062] Furthermore, the method for extracting the key performance characteristic parameters is:

[0063]

[0064] t e =t b -t a ;

[0065] Where, t e is the engagement time; s is the number of friction pairs; p is the applied pressure; R o and R i are the outer radius and inner radius of the friction part respectively; t a and t d are the start and end times of the engagement process, T(t) is the real-time friction torque, and μ is the friction coefficient COF.

[0066] Specifically, the de-noised torque data is used to extract key performance characteristic parameters, which include: friction coefficient COF and engagement time.

[0067] Furthermore, a bivariate inverse Gaussian degradation model is constructed, including:

[0068] The performance characteristic parameters are obtained through the friction torque data, and the performance characteristic parameters include the degradation process of the friction coefficient COF and the engagement time, and the probability density function of the degradation process is obtained by assuming that the random variables are controlled by the probability distribution;

[0069] By setting n clutch units to perform degradation tests under the same conditions, each unit i obtains m degradation observations. Given the latent variables of the degraded unit, the posterior distribution of the latent variables is obtained, and finally the complete log-likelihood function of the bivariate inverse Gaussian degradation model is obtained.

[0070] Specifically, the COF and bonding time data are represented as Y1(t) and Y2(t), respectively, t≥0, and they obey the inverse Gaussian distribution, that is, Y p (t)~IG(μΛ p (t), η p Λ p (t) 2), where p = 1 and 2 represent two degradation processes respectively; where μ is a common latent variable used to describe the dependency between the two degradation processes; Λ(t) is a time-dependent shape function in the form of a power law function. t, q are time and the power of time respectively; η describes the volatility of degradation; thus, the probability density function (PDF) of the degradation process can be obtained as:

[0071]

[0072] In the formula, f(Y p ) is the probability density function of the corresponding performance characteristic parameters.

[0073] As a standard practice, in this example, we assume that the random variables are governed by a probability distribution. To prevent negative values, we assume that μ -1 Obey the truncated normal distribution TN(ω,κ -2 ), and is statistically independent of the parameter η. In addition, assuming that μ -1 , the two performance parameter degradation processes Y1(t) and Y2(t) are independent of each other, so the degradation process can be expressed as Y p (t)|μ -1 ~IG(μΛ p (t),η p Λ p (t) 2 ), p = 1, 2. Therefore, the PDF of the degradation process can be rewritten as:

[0074] f(Y k (t)|μ -1 )=f(Y p );

[0075] The probability density function of the latent variable is:

[0076]

[0077] In the formula, π(μ -1 ) is the probability density function of the latent variable, ω is the mean parameter, κ is the standard deviation parameter, φ is the probability density function of the standard normal distribution, and Φ is the cumulative distribution function of the standard normal distribution.

[0078] Then, the joint PDF of the two degenerate variables is given by:

[0079] f(Y1(t),Y2(t))=∫f(Y1(t),Y2(t),μ -1 )dμ -1 ;

[0080] Where, f(Y1(t),Y2(t),μ -1)=f(Y1(t),Y2(t)μ -1 )π(μ -1 ).

[0081] Assuming that n clutch units are subjected to degradation tests under the same conditions, and each unit i obtains m degradation observations, the degradation measurement value can be expressed as:

[0082]

[0083] Among them, y p (t ij ) is the pth degradation variable of unit i at time t ij The degenerate value of k,i,j =y k (t ij )-y k (t i(j-1) ), β k,i,j =Λ k (t ij )-Λ k (t i(j-1) ).

[0084] Then, given the latent variable μ of degenerate unit i -1 , the likelihood function can be written as:

[0085]

[0086] Then, we can get the latent variable μ -1 The posterior distribution of :

[0087]

[0088] Finally, the complete log-likelihood function of the bivariate inverse Gaussian degradation model is obtained:

[0089]

[0090] Among them, θ = (ω, κ, η1, η2, q1, q2) is the parameter set of the model parameters to be estimated, and P(·) is the latent variable of a given degradation unit i. The likelihood function of the model, is the m observation data of the friction coefficient COF of unit i, is the m observation data of the joining time of unit i, is the common latent variable of the i-th clutch unit, and N is the number of clutch units.

[0091] Due to the high dimension and multiple parameters, it is difficult to directly optimize the likelihood function. Therefore, this embodiment will use a more practical Bayesian Markov Chain Monte Carlo method (MCMC) to perform parameter posterior inference and uncertainty estimation on the parameter set θ.

[0092] In order to verify the performance of the bivariate inverse Gaussian degradation model, a goodness of fit test is performed. This embodiment provides two test indicators, namely, Watanabe Akaike Information Criterion (WAIC) and Mean Absolute Error (MAE), which are calculated as follows:

[0093] WAIC = -2(LPD-k);

[0094]

[0095] Where LPD represents the sum of log-pointwise likelihoods, k represents the effective number of parameters, and Y pre and Y act are the degradation values ​​fitted by the model and the actual degradation values, respectively, and m is the number of prediction points.

[0096] The smaller the values ​​of the two indicators, the better the model performance.

[0097] Furthermore, the dynamic online prediction of clutch RUL is performed through a two-variable inverse Gaussian degradation model, including:

[0098] Use the verified degradation model to predict RUL online. According to the established degradation model and the concept of the first arrival time of equipment failure, the degradation data time window of the performance parameters is used to realize the iterative prediction of RUL of the clutch system based on the Monte Carlo simulation method. That is, through the iterative degradation data time window, the model makes the degradation inference and parameter estimation of the clutch system according to the latest degradation information, and finally obtains the predicted RUL according to the posterior sample distribution.

[0099] Specifically:

[0100] S1. Input known historical data at the current time t into the bivariate inverse Gaussian degradation model, and obtain the posterior distribution range of each parameter based on the MCMC method, including the mean and variance;

[0101] S2. When the latest data does not exceed the preset failure threshold of the key performance characteristic parameter, the latest l historical performance characteristic parameter data are selected to form a degradation data time window, and a Monte Carlo simulation method is used to generate a simulated degradation path of the key performance characteristic parameter based on the model parameter value;

[0102] S3, sampling the parameter interval in S1 S times to obtain S simulated degradation paths, and then calculating the failure time of the simulated degradation paths respectively;

[0103] S4. Obtain the remaining service life of the current clutch unit according to the estimated mean failure time and the current time.

[0104] Furthermore, the method also includes comparing the prediction performance of the two-variable inverse Gaussian degradation model with the two-variable Gamma model, respectively calculating the MAE of the two models, performing quantitative analysis on the prediction performance, and verifying the effectiveness of the two-variable inverse Gaussian degradation model.

[0105] This embodiment also provides a dual variable wet clutch online remaining life prediction system, such as Figure 2 ,include:

[0106] Sample torque storage module: used to obtain real-time friction torque data of the clutch, obtain friction torque curve, and store it in the torque storage module;

[0107] Torque preprocessing module: used to denoise and smooth the collected friction torque data, eliminate noise and abnormal points, and extract key performance characteristic parameters based on the smoothed torque data;

[0108] Bivariate Inverse Gaussian Model Module: Constructs a bivariate inverse Gaussian process model based on the processed data;

[0109] Goodness of fit test module: used to evaluate the parameter and path fitting effect estimated by the bivariate inverse Gaussian process model;

[0110] RUL Online Prediction Module: Based on the updated model and real-time data, the Monte Carlo simulation method is used to sample and predict the future trend of the degradation path, and the remaining service life of the wet clutch is dynamically calculated.

[0111] Furthermore, as another embodiment, the system also includes a performance evaluation module for testing the effectiveness and performance of the bivariate inverse Gaussian degradation model and the RUL prediction method.

[0112] In order to more clearly express the technical solution of the present invention, the following specific embodiments are provided to introduce the solution:

[0113] This embodiment provides historical data of a clutch unit for verifying a bivariate inverse Gaussian process model and a RUL prediction method.

[0114] Figure 3 and Figure 4 The torque data and data evolution diagrams of two performance characteristics obtained using the sample torque storage module and the torque preprocessing module are shown respectively.

[0115] Using the bivariate inverse Gaussian process model, the degradation data of the two performance parameters are respectively input into the bivariate inverse Gaussian model (Bi-IG) and the comparison model Gamma model (Bi-Ga), and the parameter estimation results are obtained using the MCMC method. Then, a goodness of fit test is performed based on the fitting results, and the results shown in Table 1 are obtained.

[0116] Table 1

[0117]

[0118] It can be seen from Table 1 that both indicators of the proposed inverse Gaussian model are significantly smaller than those of the classic Gamma process model, which proves its effectiveness.

[0119] Then, the RUL online prediction module is used to estimate the RUL of the clutch system at the current moment in real time based on Monte Carlo simulation, and the error index MAE is calculated. The results are as follows: Figure 5 And as shown in Table 2.

[0120] Table 2

[0121]

[0122] The method of this embodiment extracts the main performance parameters (friction coefficient and engagement time) based on the real-time friction torque data during the operation of the clutch, and accurately captures the correlation between friction plate wear and friction coefficient change in the wet clutch by constructing a two-variable inverse Gaussian (IG) degradation model, and combines Markov chain Monte Carlo (MCMC) and Monte Carlo simulation technology to achieve high-precision parameter estimation and real-time remaining life (RUL) prediction. The system proposed in this embodiment not only improves the dynamic response capability of life prediction, but also guides the formulation of maintenance decisions based on the situation, reduces the risk of sudden failures, and improves the operational reliability of the system.

[0123] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A dual variable wet clutch online remaining life prediction method, characterized in that: include: Acquiring clutch real-time friction torque data, and preprocessing the real-time friction torque data to extract key performance characteristic parameters; A two-variable inverse Gaussian degradation model is constructed based on the key performance characteristic parameters, and the dynamic online prediction of the clutch RUL is performed through the two-variable inverse Gaussian degradation model, wherein the parameters of the two-variable inverse Gaussian degradation model are estimated through the Markov chain Monte Carlo MCMC method, and the online prediction of the clutch RUL is achieved based on a large number of samplings of the Monte Carlo simulation method.

2. The method for online remaining life prediction of a dual variable wet clutch according to claim 1, characterized in that: Acquiring clutch real-time friction torque data and preprocessing the real-time friction torque data includes: The friction torque data is collected in real time by a speed torque sensor, and the friction torque data is denoised and smoothed to eliminate noise and abnormal points.

3. The method for online remaining life prediction of a dual variable wet clutch according to claim 1, characterized in that: The method for extracting the key performance characteristic parameters is: t e =t b -t a ; Where, t e It is the joining time; s is the number of friction pairs; p is the applied pressure; R o and R i are the outer radius and inner radius of the friction part respectively; t a and t d are the start and end times of the engagement process, T(t) is the real-time friction torque, and μ is the friction coefficient COF.

4. The method for online remaining life prediction of a dual variable wet clutch according to claim 3, characterized in that: Constructing the bivariate inverse Gaussian degradation model includes: Obtaining performance characteristic parameters through friction torque data, and assuming that random variables are controlled by probability distribution, thereby obtaining a probability density function of the degradation process, wherein the performance characteristic parameters include the degradation process of friction coefficient COF and engagement time; By setting n clutch units to perform degradation tests under the same conditions, each unit i obtains m degradation observations. Given the latent variables of the degraded unit, the posterior distribution of the latent variables is obtained, and finally the complete log-likelihood function of the bivariate inverse Gaussian degradation model is obtained.

5. The method for online remaining life prediction of a dual variable wet clutch according to claim 4, characterized in that: The probability density function of the degradation process is: The extracted friction coefficient COF and engagement time are set to be Y1(t) and Y2(t), t≥0, and they obey the inverse Gaussian distribution, that is, Y p (t)~IG(μΛ p (t), η p Λ p (t) 2 ), where p = 1, 2, representing two degradation processes respectively; μ is a common latent variable used to describe the dependency between the two degradation processes; Λ(t) is a time-dependent shape function in the form of a power law function. t, q are time and the power of time respectively; η describes the volatility of degradation, thereby obtaining the probability density function of the degradation process: In the formula, f(Y p ) is the probability density function of the corresponding performance characteristic parameters.

6. The method for online remaining life prediction of a dual variable wet clutch according to claim 5, characterized in that: Obtaining a probability density function of the degradation process includes: Assume that the latent variable μ -1 Obey the truncated normal distribution TN(ω,κ -2 ), and is statistically independent of the parameter η; then assuming that μ -1 , the two performance parameter degradation processes Y1(t) and Y2(t) are independent of each other, so the degradation process can be expressed as Y p (t)|μ -1 ~IG(μΛ p (t), η p Λ p (t) 2 ), p = 1, 2, so the probability density function of the degradation process is: f(Y k (t)|μ -1 )=f(Y p ); The probability density function of the latent variable is: In the formula, π(μ -1 ) is the probability density function of the latent variable, ω is the mean parameter, κ is the standard deviation parameter, φ is the probability density function of the standard normal distribution, and Φ is the cumulative distribution function of the standard normal distribution.

7. The method for online remaining life prediction of a dual variable wet clutch according to claim 6, characterized in that: The complete log-likelihood function of the bivariate inverse Gaussian degradation model is: Where θ = (ω, κ, η1, η2, q1, q2) is the parameter set of the model to be estimated, and P(·) is the latent variable of a given degradation unit i. The likelihood function of the model, is the m observation data of the friction coefficient COF of unit i, is the m observation data of the joining time of unit i, is the common latent variable of the i-th clutch unit, and N is the number of clutch units.

8. The method for online remaining life prediction of a dual variable wet clutch according to claim 1, characterized in that: The dynamic online prediction of clutch RUL is performed by the dual variable inverse Gaussian degradation model, including: Based on the Monte Carlo simulation method, the degradation data time window of the obtained performance parameters is used to realize the iterative prediction of the RUL of the clutch system, specifically: S1. Input known historical data at the current time t into the bivariate inverse Gaussian degradation model, and obtain the posterior distribution range of each parameter based on the MCMC method, including the mean and variance; S2. When the latest data does not exceed the preset failure threshold of the key performance characteristic parameter, the most recent historical performance characteristic parameter data is selected to form a degradation data time window, and a Monte Carlo simulation method is used to generate a simulated degradation path of the key performance characteristic parameter based on the model parameter value; S3, sampling the parameter interval in S1 S times to obtain S simulated degradation paths, and then calculating the failure time of the simulated degradation paths respectively; S4. Obtain the remaining service life of the current clutch unit according to the estimated mean failure time and the current time.

9. The method for online remaining life prediction of a dual variable wet clutch according to claim 1, characterized in that: The method also includes comparing the prediction performance of the two-variable inverse Gaussian degradation model with that of the two-variable Gamma model, respectively calculating the MAE of the two models, performing quantitative analysis on the prediction performance, and verifying the effectiveness of the two-variable inverse Gaussian degradation model.

10. A dual variable wet clutch online remaining life prediction system, applied to the dual variable wet clutch online remaining life prediction method according to any one of claims 1 to 9, characterized in that: include: Sample torque storage module: used to obtain real-time friction torque data of the clutch, obtain friction torque curve, and store it in the torque storage module; Torque preprocessing module: used to denoise and smooth the collected friction torque data, eliminate noise and abnormal points, and extract key performance characteristic parameters based on the smoothed torque data; Bivariate Inverse Gaussian Model Module: Constructs a bivariate inverse Gaussian process model based on the processed data; Goodness of fit test module: used to evaluate the parameter and path fitting effect estimated by the bivariate inverse Gaussian process model; RUL Online Prediction Module: Based on the updated model and real-time data, the Monte Carlo simulation method is used to sample and predict the future trend of the degradation path, and the remaining service life of the wet clutch is dynamically calculated.

Images