Automobile hub reliability analysis method combining fatigue failure and impact degradation failure

Abstract

The invention relates to the field of automobile hubs. An automobile hub reliability analysis method combining fatigue failure and impact degradation failure is characterized by comprising the following steps that firstly, fatigue failure analysis is conducted on an automobile hub, and the fatigue life is predicted, specifically, load characteristic analysis is conducted on dangerous parts of the automobile hub based on a road program stress spectrum; the fatigue life of the automobile hub under the actual use condition is predicted, and related physical characteristics and characteristic parameters of the automobile hub are obtained; then, considering the mutual influence of degradation failure and impact failure of the automobile hub, establishing a time-varying failure threshold model of the degradation failure and the impact failure according to the load characteristic analysis of the dangerous part of the automobile hub, and finally establishing an impact-degradation competitive failure reliability model based on a time-varying failure threshold in combination with the time-varying failure threshold model. And a maximum likelihood estimation method is adopted to estimate residual model parameters, and an automobile hub reliability analysis result combining fatigue failure and impact degradation failure is obtained.

Description

Technical Field

[0001] The invention relates to the field of automobile wheel hubs, and in particular to an automobile wheel hub reliability analysis method combining fatigue failure and impact-degradation competitive failure. Background Art

[0002] In actual engineering, in order to prevent huge losses caused by component failure, accurate and effective reliability analysis is of great significance. Automobile wheels are subjected to various impact loads during driving, such as vibrations from potholes and lateral forces during sharp turns. Due to the long-term influence of these variable impact loads, the wheel will gradually suffer fatigue damage, and eventually lead to fatigue failure. Wheel failure may cause the tire to fall off or burst, which in turn causes serious traffic accidents.

[0003] As an automotive component, the automotive wheel hub may experience degradation failure and impact failure during actual service, which are caused by natural degradation due to wear, aging, corrosion, temperature changes and other factors, and instantaneous random impacts caused by sudden impacts or vibrations. These two failure forms are not mutually exclusive, and the failure of the component is the result of the combined effects of degradation and impact. Considering the mutual influence of the degradation process and the impact process, the degradation failure and impact failure of the automotive wheel hub are specifically analyzed, and the reliability analysis of the automotive wheel hub is carried out by establishing an impact-degradation competition failure reliability model in consideration of the competition of the two failure forms, which is of great significance to ensure the safety of the vehicle.

[0004] In the prior art, when considering degradation failure on industrial equipment, the possible changes in the degradation failure threshold over time are usually considered. However, in the reliability prediction of automobile wheels, the main focus is on life prediction or long-term fatigue failure analysis, and no reliability analysis of specific impact-degradation competition failure is performed. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a reliability analysis method for automobile wheel hubs that combines fatigue failure and impact degradation failure, so as to solve the technical problem that the prior art does not fully consider the degradation failure and impact failure caused by the two failure processes of natural degradation and random impact that may occur in the actual service process of automobile wheel hubs, as well as the relationship between the two failure processes, and the impact on the reliability and safety of the automobile wheel hub.

[0006] Technical Solution

[0007] A reliability analysis method for automobile wheel hub combining fatigue failure and impact degradation failure, characterized by comprising the following steps:

[0008] First, fatigue failure analysis is performed on the automobile wheel hub, fatigue life is predicted, and relevant physical characteristics and characteristic parameters of the automobile wheel hub to be analyzed are obtained; specifically, based on the road program stress spectrum, load characteristics analysis is performed on the dangerous parts of the automobile wheel hub, and its fatigue life under actual use conditions is predicted, and relevant physical characteristics and characteristic parameters of the automobile wheel hub are obtained;

[0009] Then, considering the mutual influence between degradation failure and impact failure of automobile wheel hub, according to the load characteristics analysis of dangerous parts of automobile wheel hub, the time-varying failure threshold model of degradation failure and impact failure is established, and the relevant physical characteristic parameters in the model are determined;

[0010] Finally, combined with the time-varying failure threshold model, an impact-degradation competitive failure reliability model based on the time-varying failure threshold is established. The maximum likelihood estimation method is used to estimate the remaining model parameters, and the reliability analysis results of the automobile wheel hub combining fatigue failure and impact degradation failure are obtained.

[0011] Furthermore, the process of load characteristic analysis based on road program stress spectrum includes: collecting a variety of typical road pavement program stress spectra, eliminating abnormal data points and noise caused by interference factors, using mixed distribution to fit the load amplitude and mean and establish a distribution model, compiling a one-dimensional program stress spectrum and a two-dimensional program stress spectrum according to the obtained distribution model; establishing a nonlinear damage accumulation model to obtain the fatigue life of the one-dimensional program stress spectrum and the two-dimensional program stress spectrum respectively; selecting the direction with the smallest stress amplitude to strength ratio as the impact failure risk direction, and obtaining random impact strength distribution parameters according to the road load distribution characteristics for subsequent failure analysis and determination of the final model.

[0012] On the basis of the compiled road program stress spectrum, the fatigue damage under variable amplitude cyclic fatigue load is calculated, and then the stress index is introduced on the basis of the linear damage accumulation model to reflect the damage difference under stress cycles with different amplitudes, and the fatigue damage under multi-level loads with clear amplitude ratios is obtained. The nonlinear damage accumulation model is established as follows:

[0013]

[0014] Among them, σ maxi is the maximum stress of the i-th load amplitude, σ -1 is the fatigue limit stress of the material, k j is the load amplitude ratio, N fi is the stress amplitude σ of the material under the i-th load amplitude maxi Fatigue life under i is the number of load cycles of the i-th level amplitude, d1 is the damage caused by a single cycle of the 1st level load amplitude, and Ni is the fatigue life of the i-th level load amplitude.

[0015] The random impact strength distribution parameters include the impact strength mean and standard deviation.

[0016] Furthermore, through the analysis of load characteristics and damage accumulation of dangerous parts of automobile wheel hubs, the failure threshold of degradation failure gradually decreases with the passage of time. The established time-varying failure threshold model of degradation failure is: H t =H0-h t , where H0 is the initial threshold and h is the threshold time-varying coefficient.

[0017] Through the analysis of load characteristics and damage accumulation of dangerous parts of automobile wheel hubs, the failure threshold of impact failure gradually decreases with the passage of time. The established time-varying failure threshold model of impact failure is: t =D0-dt, where D0 is the initial threshold and d is the time-varying coefficient of the threshold.

[0018] Furthermore, the impact-degradation competition failure reliability model based on time-varying failure threshold is:

[0019]

[0020] Among them, α and β are degradation process parameters, μ w and σ w is the random impact strength distribution parameter, λ is the random impact number distribution parameter, κ is the conversion coefficient between impact strength and degradation, H t and D t are the soft and hard time-varying failure thresholds respectively.

[0021] The random impact strength distribution parameters and are obtained from the pavement load distribution characteristics, which are the impact strength mean and standard deviation respectively. The impact strength conversion coefficient is obtained by converting the life and damage, and the conversion coefficient is determined by selecting the medium damage amount.

[0022] When estimating the parameters of the reliability functions α, β and λ, their probability density distribution functions are first obtained, and then the maximum likelihood function is solved based on the observed data to obtain the maximum likelihood estimates of each parameter.

[0023] Beneficial Effects

[0024] The present invention performs fatigue failure analysis on automobile wheel hubs, predicts fatigue life, obtains relevant physical properties and characteristic parameters of the automobile wheel hub to be analyzed, then considers the mutual influence of degradation failure and impact failure of the automobile wheel hub, and establishes a time-varying failure threshold model of degradation failure and impact failure according to the load characteristic analysis of dangerous parts of the automobile wheel hub. Finally, in combination with the time-varying failure threshold model, an impact-degradation competitive failure reliability model based on the time-varying failure threshold is established to obtain the reliability analysis result of the automobile wheel hub combining fatigue failure and impact degradation failure. Compared with the current situation of simply performing reliability analysis based on fatigue life, the reliability analysis is closer to the actual operation situation, and the reliability analysis of the automobile wheel hub is more accurate by organically combining the two analyses, which is very helpful for automobile design. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0026] Figure 2 Schematic diagram of fatal and non-fatal impacts on the wheel hub at different times.

[0027] Figure 3 Schematic diagram of the multi-impact failure model.

[0028] Figure 4 Schematic diagram of reliability changes for three failure modes: variable threshold competition failure, fixed threshold competition failure, and degradation failure. DETAILED DESCRIPTION

[0029] The present invention is further described below in conjunction with specific embodiments and drawings.

[0030] In order to solve the problem that the prior art does not fully consider the degradation failure and impact failure caused by the two failure processes of natural degradation and random impact that may occur in the actual service of automobile wheels, the present application proposes to first analyze the load characteristics of dangerous parts of automobile wheels based on the road program stress spectrum, and predict its fatigue life under actual use conditions, so as to provide a basis for the subsequent model parameter determination; then, because the automobile wheel may experience degradation failure and impact failure in the actual service process, considering the mutual influence of the degradation process and the impact process, a time-varying failure threshold model is established, and the model parameters are determined according to the actual physical meaning; finally, considering the competitive dependence of degradation failure and impact failure, an impact-degradation competitive failure reliability model is established, and the maximum likelihood estimation method is used to estimate the remaining parameters.

[0031] Firstly, a six-component force meter is used to collect the programmed stress spectra of various typical road surfaces such as serpentine cobblestone roads, gravel roads, and washboard roads in the test field. During the test, the vehicle speed is kept stable and consistent with the daily driving speed.

[0032] The collected initial load data contains abnormal data points and noise caused by interference factors. Before compiling the program stress spectrum, the signal needs to be preprocessed by a specific processing method. The super-threshold extrapolation method is used to extrapolate the road load by 10 times, and then the peak and valley values ​​of the extrapolated signal are extracted and counted. The rain flow counting method is used to count and statistically analyze the load-time history cycles of the peak and valley values. The load amplitude and mean are fitted using a mixed distribution and a distribution model is established. The one-dimensional program stress spectrum and the two-dimensional program stress spectrum are compiled based on the obtained distribution model.

[0033] When compiling a one-dimensional program stress spectrum, the measured random load amplitude needs to be divided into multiple levels, and the frequency of each level of amplitude needs to be counted. The present invention uses the amplitude ratio method to divide the program stress spectrum into 8 levels. The amplitude division interval is determined by formula (1).

[0034]

[0035] In the formula, A min is the minimum stress amplitude; A max is the maximum stress amplitude; l i and l i-1 are the upper and lower limits of the i-th level load amplitude range respectively; k i is the proportional coefficient of the i-th level amplitude. When the load is graded into 8 levels, the specific values ​​of the load amplitude ratios of each level are shown in Table 1.

[0036] Table 1 Proportional coefficients of load amplitudes at various levels

[0037]

[0038] Under variable amplitude cyclic fatigue load, the fatigue damage caused after n cycles is:

[0039]

[0040] Where n i is the number of cycles of the i-th load amplitude, D is the total fatigue damage caused after n cycles, N i is the fatigue life of the i-th load amplitude.

[0041] A nonlinear damage accumulation model is established on the basis of the linear model, and a stress index is introduced on the basis of the linear damage accumulation model to reflect the damage difference under stress cycles of different amplitudes, as shown in formula (3).

[0042]

[0043] In the formula, q i is the stress index, which can be determined by formula (4).

[0044]

[0045] Among them, σ maxi is the maximum stress of the i-th load amplitude, σ -1 is the fatigue limit stress of the material. Under the same amplitude stress cycle, the damage accumulation of the parts is linear. According to formula (2),

[0046]

[0047] Where di is the damage caused by a single cycle of the i-th load amplitude, d i+1 is the damage caused by a single cycle of the i+1th load amplitude, N fi is the stress amplitude σ of the material under the i-th level load maxi Fatigue life under f(i+1) is the stress amplitude σ of the material under the i+1th level load max(i+1) Fatigue life under. Multi-level loads are usually expressed as program stress spectra. The amplitude ratio of each level of load is known. The difference in damage of each level of load is expressed by the load amplitude ratio, and then:

[0048]

[0049] In the formula, k i is the proportional coefficient of the i-th level amplitude, k j is the load amplitude ratio, and d1 is the damage caused by a single cycle of the first level load amplitude.

[0050] Then the fatigue damage under multi-level loads with a clear amplitude ratio can be expressed as:

[0051]

[0052] During operation, the wheel hub is always under six-way multiaxial load. When calculating the fatigue life of each level of the one-dimensional program stress spectrum, it is calculated in matrix form, and the six-way stress is included. The calculated number of cycles and damage under each level of load are shown in Table 2.

[0053] Table 2 Load cycle number and fatigue damage

[0054]

[0055]

[0056] According to the nonlinear damage accumulation model of formula (7), the single damage degree of the cyclic one-dimensional program stress spectrum is 1.80859×10 -3 , the fatigue life of the one-dimensional program stress spectrum is 553 times.

[0057] The two-dimensional program stress spectrum has graded both the load amplitude and the mean value. When using the progressive damage model to estimate the fatigue life corresponding to the two-dimensional stress spectrum, the influence of the load mean value on the fatigue life needs to be converted to the equivalent stress amplitude for calculation.

[0058] The commonly used method for calculating equivalent stress amplitude in engineering is the Goodman model, as shown in formula (8).

[0059]

[0060] In the formula, σ eq is the equivalent stress amplitude; σ a is the stress amplitude; σ m is the stress mean; σ u is the ultimate tensile strength.

[0061] The Goodman model is improved. When the amplitude under the positive mean is equivalent, the tensile strength is used for correction; when the amplitude under the negative mean is equivalent, the compressive strength is used for correction. The equivalent stress amplitude correction formula is as follows:

[0062]

[0063] In the formula, σ ut is the tensile strength; uc is the compressive strength; a, b are constants with values ​​of 0 or 1 and a+b=0. m When is positive, a=1, otherwise b=1.

[0064] Considering the extreme failure of random impact, the direction with the smallest ratio of stress amplitude to strength is selected as the impact failure danger direction. Table 3 shows the corrected two-dimensional program stress spectrum in the danger direction. It can be seen from the table that under the influence of the mean, the actual amplitude of the damage is greater than the amplitude obtained by the cycle count.

[0065] Table 3 Corrected two-dimensional program stress spectrum in the dangerous direction

[0066]

[0067] Similar to the one-dimensional program stress spectrum life estimation, when calculating the two-dimensional program stress spectrum life, the loads at each level in each direction are corresponded to form a multi-axis program stress spectrum for calculation. The corresponding life of each level of load is shown in Table 4.

[0068] Table 4 Number of load cycles at each level

[0069]

[0070] According to the load occurrence frequency and the linear damage accumulation criterion, the fatigue damage corresponding to each level of load is shown in Table 5.

[0071] Table 5 Fatigue damage at various load levels

[0072]

[0073] According to the nonlinear damage accumulation model of the previous formula (7), the damage degree of the two-dimensional program stress spectrum is 2.29704×10 -3 , the fatigue life is 536 times.

[0074] Then, the reliability analysis of hub degradation failure and impact failure is carried out, and the impact-degradation competition failure reliability model is established considering the competition between the two failure forms.

[0075] According to the variation characteristics of degradation amount, Wiener process is used to model the degradation process.

[0076] The degradation process expression based on the Wiener model is:

[0077] X(t)=X(0)+αt+βB(t) (10)

[0078] In the formula, X(0) represents the initial performance state of the system, X(0) = 0; α is the drift coefficient of the degradation process, which controls the degradation rate of the components; β is the degradation diffusion coefficient, which controls the degree of fluctuation of the degradation process; B(t) represents the standard Brownian motion, which reflects the randomness and uncertainty of the degradation process. It can be seen that the expectation of X(t) E(X(t)) = αt, and the variance of X(t) S(X(t)) = β 2 t, that is, X(t)~N(αt,β 2 t). N represents the symbol of normal distribution, which means that X(t) follows a normal distribution with αt as mean and β as value. 2 t is a normal distribution with variance.

[0079] When the degradation state of a component exceeds the set failure threshold H, it is considered that degradation failure has occurred. The component degradation failure life distribution T = inf(t|X(t)≥H), and the probability density function of the Wiener degradation process system life T is:

[0080]

[0081] make The probability density function of the system lifetime T can be rewritten as

[0082]

[0083] From formula (12), we can see that the system life has a mean value of H / α and a shape parameter of H 2 / β 2 The inverse Gaussian distribution of Then the cumulative distribution function of component degradation life is

[0084]

[0085] According to the relationship between the various characteristic quantities of reliability, we can know that the reliability R T (t) = 1-F T (t), then the degradation life reliability is

[0086]

[0087] Wherein, Φ(·) in formula (13) and formula (14) represents the cumulative distribution function of the standard normal distribution.

[0088] When considering the impact of external shock on reliability, we mainly start from three aspects: shock intensity, shock law and shock failure type. Low-intensity shock will not cause component failure, but it will cause certain damage or performance degradation; high-intensity shock will cause greater damage or degradation, and is more likely to cause component failure than low-intensity shock. However, it is not possible to judge whether the shock directly causes component failure based on the shock intensity alone, and it should be analyzed in combination with the shock failure type.

[0089] Starting from the common random impact distribution and impact failure types, the impact failure of components is analyzed in detail.

[0090] According to the actual situation of the shock, the homogeneous Poisson process {N(t), t≥0} with parameter λ is used to describe the external random shock process of the parts, where N(t) represents the total number of random shocks arriving from 0 to t. From a physical sense, it can be seen that the total number of random shocks N(t) is a non-negative random variable. In any time period of length τ, the total number of shocks n arriving at the system obeys the homogeneous Poisson distribution with parameter λτ, and its probability function can be expressed as:

[0091]

[0092] According to the operation of parts in normal working process, the random impacts that come can be divided into non-lethal impact and lethal impact according to the degree of damage they cause. Non-lethal impact is the impact whose impact intensity is lower than the impact failure threshold. When the parts are hit by such impact, they will not fail immediately, but will increase some damage or performance degradation; lethal impact is the impact whose impact intensity is higher than the impact failure threshold. When the parts are hit by such impact, the impact failure will occur immediately. In actual service, most of the impacts are non-lethal impacts, such as the attached Figure 2As shown in the figure, W1, W2, W3, W4, and W5 are the external impacts suffered at t1, t2, t3, t4, and t5 respectively, and D is the impact failure threshold. It can be seen from the figure that the part suffered a fatal impact at t5 and suffered impact failure, while W1, W2, W3, and W4 are all non-fatal impacts.

[0093] The number of impact failure models assumes that failure occurs when the total number of impacts exceeding a certain critical value reaches a certain number, such as Figure 3 As shown, D is the hard failure threshold and M is the critical value. For example, some vehicles with good sports performance have a set number of ejection starts when they leave the factory. When the actual number of ejections exceeds the set number, the vehicle can no longer be ejected.

[0094] Combined with damage accumulation, a time-varying failure threshold is introduced in reliability modeling, and a competitive failure reliability model is established when considering the mutual influence of degradation failure and impact failure.

[0095] Different failure modes have different failure thresholds. The failure threshold of degradation failure is the soft failure threshold, denoted as H, and the failure threshold of impact failure is the hard failure threshold, denoted as D. When considering degradation failure or impact failure separately, the respective failure thresholds are usually considered as constants for analysis. However, in actual engineering, the soft and hard failure thresholds will change due to cyclic loads or impact loads. This change is generally time-varying, and in most cases, the threshold gradually decreases with time.

[0096] According to the previous study on damage accumulation, it is found that the soft and hard failure thresholds gradually decrease over time. Therefore, a linear model is used to describe the changes in the soft and hard thresholds, as shown in Equations (16) and (17).

[0097] H t =H0-ht (16)

[0098] D t =D0-dt (17)

[0099] Where H0 is the initial soft threshold, h is the time-varying coefficient of the soft threshold; D0 is the initial hard threshold, d is the time-varying coefficient of the hard threshold.

[0100] During the load-bearing process, the hub stiffness continuously degrades. The degradation failure process is regarded as the process of stiffness degradation to failure. The impact failure is the failure that occurs when the impact strength is higher than the residual strength. The time-varying soft failure threshold is related to the stiffness degradation, and the time-varying hard failure threshold is related to the strength degradation. Assume that at the time of degradation failure, the initial stiffness is 3789MPa, and the ratio of failure stiffness to initial stiffness is 0.14; at the time of impact failure, the initial strength is 1140Mpa, and the ratio of failure strength to initial strength is 0.4. The life and time are converted to a duration of 299496.7h, and the specific forms of the time-varying soft and hard failure thresholds can be obtained, as shown in equations (18) and (19).

[0101] H t =3789-0.012145t (18)

[0102] D t =1140-0.002284t (19)

[0103] According to the previous study of road loads, it can be found that the actual working load of the wheel hub is composed of low-amplitude and low-mean cyclic loads and high-amplitude and high-mean impact loads. The following assumptions are made for the model:

[0104] (1) Before failure, the wheel hub undergoes performance degradation and random impact processes;

[0105] (2) The performance degradation amount is recorded as X(t). When X(t) exceeds the soft failure threshold H(t) for the first time, degradation failure occurs.

[0106] (3) The impacts on the wheel hub are divided into two categories: fatal impacts exceeding the hard failure threshold D(t) and non-fatal impacts not exceeding the hard failure threshold D(t). When the wheel hub suffers a fatal impact, it is considered as an impact failure. When the wheel hub suffers a non-fatal impact, the sudden increase in degradation caused by each impact is included in the total degradation.

[0107] Note X S (t) is the total degradation at time t,

[0108] X S (t) = X (t) + S (t) (20)

[0109] Among them, X(t) is the degradation amount of the degradation process, and S(t) is the degradation amount caused by non-fatal impact in the impact process.

[0110] This scheme uses the Wiener degradation model to analyze the degradation process and the homogeneous Poisson model to analyze the random impact process. According to the previous research on pavement load, the impact intensity obeys the normal distribution. Assume that the number of impacts N(t) in the random impact process obeys the homogeneous Poisson distribution with parameter λt, that is, N(t)~P(λt), where the number of non-fatal impacts and the number of fatal impacts are N1(t) and N2(t), respectively, and the probability of occurrence is p1 and p2, respectively. The intensity of each impact is recorded as w and That is, the intensity W of each impact obeys a mean value μ w , the variance is σ w 2 The normal distribution of the part is recorded as Y for each non-fatal impact, and the degradation amount and impact intensity satisfy Y i =κW i , κ is the conversion factor, and the degradation amount Y also obeys the normal distribution. Then:

[0111]

[0112] In formula (21), D(t) is the time-varying value of the hard failure threshold D. If no failure occurs within time t, there are two cases: there is no random shock within time t, only degradation process, and no degradation failure occurs; there is a random shock within time t, but no fatal shock, and no shock failure or degradation failure occurs.

[0113] The reliability of the first case is:

[0114]

[0115] The reliability of the second case is:

[0116]

[0117] During the entire working cycle of the hub, whenever it is in the above two situations, these two situations are independent and mutually exclusive events. The reliability of the hub at time t is:

[0118]

[0119] The maximum likelihood estimation method estimates model parameters by maximizing the likelihood function of observed data. It has good consistency and intuitiveness. According to the characteristics of the reliability function, this paper uses the maximum likelihood estimation method for parameter estimation.

[0120] There are many unknown parameters in the reliability function, and the specific values ​​need to be determined by parameter estimation method. Among them, α and β are degradation process parameters, μ w and σ wis the random impact strength distribution parameter, λ is the random impact number distribution parameter, κ is the conversion coefficient between impact strength and degradation, H t and D t They are soft and hard time-varying failure thresholds, respectively. Due to the complex form of the reliability function, the maximum likelihood estimation method is complex and computationally intensive when used to estimate all parameters. Before performing parameter estimation, the parameter values ​​with actual physical significance are determined based on the physical model, which can both ensure the accuracy of the model and reduce the amount of computation.

[0121] Among the unknown parameters, those with actual physical meaning are: μ w , σ w , H t , D t and κ. Among them, the random impact strength distribution parameter μ w and σ w It can be obtained from the pavement load distribution characteristics. Considering the failure of random impact extreme value, the direction with the smallest ratio of stress amplitude to strength is selected as the impact failure risk direction. The load mean follows the normal distribution. Considering the influence of amplitude, the impact strength mean μ can be obtained. w =4.754MPa, standard deviation σ w =42.76MPa, the impact strength conversion coefficient κ can be obtained by converting the life and damage. In order to reflect the surge in degradation caused by random impact and ensure that the reliability estimate is not too conservative, the medium damage amount is used to determine the conversion coefficient κ, then κ=1×10 -5 ; Soft and hard time-varying failure threshold H t , D t This has been explained before.

[0122] Normally, when estimating the parameters of the reliability function, one should first obtain its probability density distribution function, and then solve the maximum likelihood function based on the observed data to obtain the maximum likelihood estimates of each parameter. However, in the shock-degradation competition failure process, the observed data cannot reflect the occurrence of random shocks. Parameter estimation is divided into two parts.

[0123] The degradation process parameters are estimated using the degradation data samples measured by the same components under non-impact tests. Assuming there are n samples, the life of the i-th sample at failure is T i , considering the existence of a degradation threshold in the system, the likelihood function is set according to the probability density function of the lifetime

[0124]

[0125] Taking the logarithm of both sides of the likelihood function, we get the log-likelihood function

[0126]

[0127] Find the partial derivative of the estimated parameter

[0128]

[0129] The maximum likelihood estimates of the drift and diffusion coefficients are

[0130]

[0131]

[0132] The impact load data samples are used to estimate the parameters of the impact process. Assuming there are n samples, the i-th sample has N occurrences within the data collection time. i random shocks, setting the likelihood function according to the Poisson distribution

[0133] Taking the logarithm of both sides of the likelihood function, we have:

[0134]

[0135] Take the derivative of the estimated parameter and set the derivative to 0. The maximum likelihood estimate of the parameter λ can be obtained by solving

[0136]

[0137] In this way, all parameters have been obtained. The reliability analysis application of the method of the present invention is carried out. Taking the program stress spectrum of the test track road surface as an example, the reliability of variable threshold competition failure, fixed threshold competition failure and degradation failure is obtained by table lookup and calculation, as shown in the attached figure. Figure 4 As shown in the figure, the reliability changes of the three failure modes all show a trend of first slowly decreasing, then decreasing at a uniform rate, and finally decreasing rapidly. Compared with competitive failure, the reliability of the degradation failure process changes slowly, and the difference between the three stages is not obvious. The reliability reduction rate is approximately constant, which reflects the characteristic of stable degradation of hub performance when there is no random impact. When the use time is short, the impact of random impact is small, and there may even be no random impact, so the reliability of the three failure modes decreases at a similar rate, and the reliability curves have a high degree of overlap. After a certain period of use, the appearance of random impact makes the reliability of degradation failure and competitive failure significantly different.

[0138] In the case of competitive failure, the reliability of the time-varying threshold decreases faster than the fixed threshold, and the difference in reliability decreases becomes more and more obvious as time goes by. This is because the failure threshold gradually decreases over time, making the hub more likely to fail under the same load condition. The 95% reliability life, 50% reliability life, and 10% reliability life under each failure mode are shown in Table 6.

[0139] Table 6 Reliability life under different failure modes

[0140]

[0141] It can be seen from Table 6 that the 95% reliability life of the three failure modes differs by about 15%, the 50% reliability life of the variable threshold and the fixed threshold differs by about 7%, and the 10% reliability life differs by about 10%. When designing the wheel hub, if the wheel hub is to have a higher reliability, the variable threshold competition failure should be considered. It has been verified that the above method meets the expected requirements and the present invention has accuracy.

[0142] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A reliability analysis method for automobile wheels combining fatigue failure and impact degradation failure, characterized in that The steps include: First, fatigue failure analysis is performed on the automobile wheel hub, fatigue life is predicted, and relevant physical characteristics and characteristic parameters of the automobile wheel hub to be analyzed are obtained; specifically, based on the road program stress spectrum, load characteristics analysis is performed on the dangerous parts of the automobile wheel hub, and its fatigue life under actual use conditions is predicted, and relevant physical characteristics and characteristic parameters of the automobile wheel hub are obtained; Then, considering the mutual influence between degradation failure and impact failure of automobile wheel hub, according to the load characteristics analysis of dangerous parts of automobile wheel hub, the time-varying failure threshold model of degradation failure and impact failure is established, and the relevant physical characteristic parameters in the model are determined; Finally, combined with the time-varying failure threshold model, an impact-degradation competitive failure reliability model based on the time-varying failure threshold is established. The maximum likelihood estimation method is used to estimate the remaining model parameters, and the reliability analysis results of the automobile wheel hub combining fatigue failure and impact degradation failure are obtained.

2. The automobile wheel hub reliability analysis method according to claim 1, characterized in that: The process of load characteristic analysis based on road program stress spectrum includes: collecting multiple typical road pavement program stress spectra, eliminating abnormal data points and noise caused by interference factors, using mixed distribution to fit the load amplitude and mean and establish a distribution model, compiling a one-dimensional program stress spectrum and a two-dimensional program stress spectrum according to the obtained distribution model; establishing a nonlinear damage accumulation model to obtain the fatigue life of the one-dimensional program stress spectrum and the two-dimensional program stress spectrum respectively; selecting the direction with the smallest stress amplitude to strength ratio as the impact failure risk direction, and obtaining random impact strength distribution parameters according to the road load distribution characteristics for subsequent failure analysis and determination of the final model.

3. The automobile wheel hub reliability analysis method according to claim 2, characterized in that: On the basis of the compiled road program stress spectrum, the fatigue damage under variable amplitude cyclic fatigue load is calculated, and then the stress index is introduced on the basis of the linear damage accumulation model to reflect the damage difference under stress cycles with different amplitudes, and the fatigue damage under multi-level loads with clear amplitude ratios is obtained. The nonlinear damage accumulation model is established as follows: Among them, σ maxi is the maximum stress of the i-th load amplitude, σ -1 is the fatigue limit stress of the material, k j is the load amplitude ratio, N fi is the stress amplitude σ of the material under the i-th load amplitude maxi Fatigue life under i is the number of load cycles of the i-th level amplitude, d1 is the damage caused by a single cycle of the 1st level load amplitude, and Ni is the fatigue life of the i-th level load amplitude.

4. The automobile wheel hub reliability analysis method according to claim 2, characterized in that: The random impact strength distribution parameters include the impact strength mean and standard deviation.

5. The automobile wheel hub reliability analysis method according to claim 1, characterized in that: Through the analysis of load characteristics and damage accumulation of dangerous parts of automobile wheel hubs, the failure threshold of degradation failure gradually decreases with the passage of time. The established time-varying failure threshold model of degradation failure is: H t =H0-ht, where H0 is the initial threshold and h is the time-varying coefficient of the threshold.

6. The automobile wheel hub reliability analysis method according to claim 5, characterized in that: Through the analysis of load characteristics and damage accumulation of dangerous parts of automobile wheel hubs, the failure threshold of impact failure gradually decreases with the passage of time. The established time-varying failure threshold model of impact failure is: t =D0-dt, where D0 is the initial threshold and d is the time-varying coefficient of the threshold.

7. The automobile wheel hub reliability analysis method according to claim 1, characterized in that: The impact-degradation competition failure reliability model based on time-varying failure threshold is: Among them, α and β are degradation process parameters, μ w and σ w is the random impact strength distribution parameter, λ is the random impact number distribution parameter, κ is the conversion coefficient between impact strength and degradation, H t and D t are the soft and hard time-varying failure thresholds respectively.

8. The automobile wheel hub reliability analysis method according to claim 7, characterized in that: The random impact strength distribution parameter μ w and σ w The impact strength conversion coefficient κ is obtained by converting the life and damage, and the medium damage amount is selected to determine the conversion coefficient.

9. The automobile wheel hub reliability analysis method according to claim 8, characterized in that: When estimating the parameters of the reliability functions α, β and λ, their probability density distribution functions are first obtained, and then the maximum likelihood function is solved based on the observed data to obtain the maximum likelihood estimates of each parameter.

Images