A method and system for structural fatigue damage assessment and residual life prediction

Abstract

The application discloses a kind of structural member fatigue damage evaluation, residual life prediction method and system, according to fatigue damage distribution calculation result and nonlinear ultrasonic detection damage value, based on the principle of probability maximization obtains correction fatigue damage conditional probability, according to correction fatigue damage conditional probability to fatigue damage distribution and nonlinear ultrasonic detection fatigue damage distribution are fused, and correction fatigue damage distribution is obtained, according to correction fatigue damage distribution calculation structural member current fatigue damage value.Based on the fatigue damage threshold of structural member and current fatigue damage value obtains the residual life distribution function of structural member, according to residual life distribution function calculation structural member's residual life.The application provides a kind of structural member fatigue damage evaluation, residual life prediction method and system, realize structural member residual life prediction, according to the maintenance decision of early structural member made according to prediction result, guarantee structural member safe and reliable operation.

Description

Technical Field

[0001] This invention relates to a method and system for assessing fatigue damage and predicting remaining life of structural components, belonging to the field of structural component degradation reliability technology. Background Technology

[0002] Fatigue failure is the predominant form of structural failure in engineering machinery and is a typical type of degradation failure. Structural components are subjected to alternating cyclic loads. When fatigue damage reaches a certain threshold, fatigue failure occurs, which may lead to major safety accidents, causing casualties and huge economic losses. Therefore, it is necessary to assess the degree of damage to structural components in real time in order to make maintenance decisions in advance and take measures to ensure safety.

[0003] An existing fatigue testing method and remaining life prediction method for metal components rely solely on nonlinear ultrasonic coefficients to determine damage to the metal components, without considering the actual load history of the metal components. This method can only provide fatigue damage warning to a certain extent and cannot achieve real-time and accurate evaluation of structural components.

[0004] Existing fatigue life prediction methods, based on fatigue crack propagation models, require testing multiple samples to obtain fatigue crack propagation curves. The evaluation results represent the overall fatigue life of the structural component, without considering individual differences, and cannot assess the fatigue damage of individual structural components in real time. Consequently, the evaluation results deviate from reality.

[0005] Therefore, those skilled in the art urgently need to solve the technical problems existing in the prior art for assessing the degradation of structural components and predicting their remaining life. Summary of the Invention

[0006] Objective: To overcome the shortcomings of existing technologies, this invention provides a method and system for assessing fatigue damage and predicting the remaining life of structural components. It performs degradation modeling based on individual fatigue degradation processes to predict the remaining life of structural components. Based on the prediction results, maintenance decisions can be made in advance to ensure the safe and reliable operation of structural components.

[0007] Technical solution: To solve the above technical problems, the technical solution adopted by the present invention is as follows:

[0008] A method for assessing fatigue damage and predicting remaining life of structural components includes the following steps:

[0009] Step 1: Obtain the stress amplitude in each stress cycle of the structural component, fit the probability density function of the stress amplitude based on the stress amplitude, and obtain the fatigue damage calculation formula based on the stress amplitude and the probability density function of the stress amplitude.

[0010] Step 2: Based on the fatigue degradation process of the structural component, obtain the Wiener process model, calculate the fatigue damage of the structural component at each time step according to the fatigue damage calculation formula, calculate the parameters of the Wiener process model based on the fatigue damage of the structural component at each time step, and obtain the fatigue damage distribution based on the parameters.

[0011] Step 3: Obtain the nonlinear ultrasonic coefficient of the structural component, and establish the fatigue damage distribution of nonlinear ultrasonic testing based on the relationship between the nonlinear ultrasonic coefficient and the nonlinear ultrasonic detection damage value.

[0012] Step 4: Based on the fatigue damage distribution calculation results and the nonlinear ultrasonic detection damage value, obtain the corrected fatigue damage conditional probability based on the probability maximization principle. Then, fuse the fatigue damage distribution and the nonlinear ultrasonic detection fatigue damage distribution according to the corrected fatigue damage conditional probability to obtain the corrected fatigue damage distribution. Finally, calculate the current fatigue damage value of the structural component based on the corrected fatigue damage distribution.

[0013] As a preferred embodiment, the method further includes: Step 5: Obtain the remaining life distribution function of the structural component based on the fatigue damage threshold and the current fatigue damage value, and calculate the remaining life of the structural component based on the remaining life distribution function.

[0014] As a preferred embodiment, step 1 includes:

[0015] The stress amplitude s and the probability density function p(s) of the stress amplitude in each stress cycle of the structural component are obtained, and the fatigue damage calculation formula is as follows:

[0016]

[0017] Where D(t) represents the fatigue damage caused by all stress cycles at time t, k and c are material constants, f0 is the stress return-to-zero rate per unit time, and ds is a small increment of the stress amplitude.

[0018] As a preferred embodiment, step 2 includes:

[0019] Based on the fatigue degradation process of the structural components, a Wiener process model is obtained, and the calculation formula for the Wiener process model is as follows:

[0020] D(t)=D(0)+μt+σB(t)

[0021] Where D(t) represents the fatigue damage caused by all stress cycles at time t, D(0) represents the initial degradation state of the structural component, μ is the drift coefficient, σ is the diffusion coefficient, and B(t) represents the standard Wiener process.

[0022] The fatigue damage D at time t is calculated using the fatigue damage calculation formula. t =D(t=t)i ), t i The value range is from the 0th time to the rth time, i = 0, 1, 2, ..., r.

[0023] Let ΔD t =D t -D t-1 For time t i-1 By time t i The degradation increment, Δt i =t i -t i-1 .

[0024] The parameters of the Wiener process model are calculated using the following formula:

[0025]

[0026] According to D t =D t-1 +μ t The fatigue damage distribution is obtained, and the formula for calculating the fatigue damage distribution is as follows:

[0027] D t |D t-1 ,μ t ~N(μ1,σ1) 2 )=N(μt,σ 2 t)

[0028] Where, μ t Let μ1 be the damage rate at time t, and μ1 = μt, σ1 2 =σ 2 t.

[0029] As a preferred embodiment, step 3 includes:

[0030] The nonlinear ultrasonic coefficient β of the structural component at time t is obtained. t According to the nonlinear ultrasonic coefficient β t Calculation of damage value Z in nonlinear ultrasonic testing t Z t The calculation formula is as follows:

[0031] Z t =g(β) t )

[0032] Wherein, g(β) t ) is the nonlinear ultrasonic coefficient β t The relationship between fatigue and damage is characterized.

[0033] Based on the nonlinear ultrasonic detection damage value Z t The fatigue damage distribution detected by nonlinear ultrasonic testing was established, and the calculation formula for the fatigue damage distribution detected by nonlinear ultrasonic testing is as follows:

[0034] Z t |D t ~N(μ2,σ2) 2 )

[0035] Where, μ2=E(Z) t )=E(g(β t )), σ2 2 =Var(Z) t )=Var(g(β t )).

[0036] As a preferred embodiment, step 4 includes:

[0037] The fatigue damage state D at times t and t-1 is calculated based on the fatigue damage distribution. t D t-1 Obtain the nonlinear ultrasonic damage value Z at time t. t Based on the principle of probability maximization, the corrected fatigue damage conditional probability is obtained. The formula for calculating the corrected fatigue damage conditional probability is as follows:

[0038] P(D t |D t-1 Z t ,μ t )∝P(Z t |D t )P(D t |D t-1 ,μ t )

[0039] Where, μ t Let t be the damage rate at time t.

[0040] The fatigue damage distribution is obtained by fusing the fatigue damage distribution with the fatigue damage distribution detected by nonlinear ultrasonic testing based on the modified fatigue damage conditional probability. The formula for calculating the modified fatigue damage distribution is as follows:

[0041]

[0042] in,

[0043] The current fatigue damage value of the structural component is calculated based on the corrected fatigue damage distribution.

[0044] As a preferred embodiment, step 5 includes:

[0045] Based on the current fatigue damage value of the structural component and the failure threshold L of the structural component, the remaining life distribution function of the structural component is obtained. The formula for calculating the remaining life distribution function of the structural component is as follows:

[0046]

[0047] Where μ is the drift coefficient, σ is the diffusion coefficient, and t is time. The remaining lifetime is obtained based on the remaining lifetime distribution function of the structural component. The remaining lifetime includes the average remaining lifetime and the 95% one-sided confidence lower limit of the remaining lifetime.

[0048] The formula for calculating the average remaining life is as follows:

[0049]

[0050] The formula for calculating the 95% one-sided confidence lower limit of the remaining lifetime is as follows:

[0051]

[0052] Secondly, a fatigue damage assessment and remaining life prediction system for structural components includes the following modules: a load acquisition module, used to acquire the stress amplitude in each stress cycle of the structural component.

[0053] Cyclic counting and distribution fitting module: used to fit the probability density function of stress amplitude based on stress amplitude, and obtain the fatigue damage calculation formula based on stress amplitude and the probability density function of stress amplitude.

[0054] Fatigue damage calculation module: used to obtain the Wiener process model based on the fatigue degradation process of structural components, and obtain the fatigue damage distribution.

[0055] Fatigue Degradation Modeling Module: This module calculates the fatigue damage of structural components at each time step based on the fatigue damage calculation formula, calculates the parameters of the Wiener process model based on the fatigue damage of structural components at each time step, and sends the parameters to the fatigue damage calculation module.

[0056] Nonlinear ultrasonic testing coefficient measurement module: used to determine the nonlinear ultrasonic coefficient of structural components and obtain the nonlinear ultrasonic coefficient of structural components.

[0057] Fatigue damage conversion module: used to establish the fatigue damage distribution of nonlinear ultrasonic detection based on the relationship between the nonlinear ultrasonic coefficient and the nonlinear ultrasonic detection damage value.

[0058] Fatigue damage update module: Based on the fatigue damage distribution calculation results and the nonlinear ultrasonic detection damage value, it obtains the corrected fatigue damage conditional probability based on the probability maximization principle, fuses the fatigue damage distribution and the nonlinear ultrasonic detection fatigue damage distribution according to the corrected fatigue damage conditional probability to obtain the corrected fatigue damage distribution, and calculates the current fatigue damage value of the structural component according to the corrected fatigue damage distribution.

[0059] The remaining life prediction module is used to obtain the remaining life distribution function of the structural component based on the fatigue damage threshold and the current fatigue damage value, and to calculate the remaining life of the structural component based on the remaining life distribution function.

[0060] As a preferred option, it also includes: a maintenance decision module: used to formulate structural component maintenance / replacement plans in advance based on the remaining life prediction results.

[0061] Beneficial effects: The fatigue damage assessment and remaining life prediction method and system for structural components provided by this invention have the following advantages over the prior art:

[0062] (1) The present invention proposes a real-time fatigue damage calculation method based on the time history of pulsating load, which is easy to implement on a computer and requires little computation and storage.

[0063] (2) The method proposed in this invention makes full use of load history information and detection results, which can improve the evaluation accuracy and make the results more accurate.

[0064] (3) The fatigue degradation process established by this invention can predict the remaining life of structural components, provide a basis for decision-making on the maintenance of structural components, ensure the safety of structural components, and improve the utilization rate. Attached Figure Description

[0065] Figure 1 This is a schematic diagram of the stress load history of a structural component.

[0066] Figure 2 This is a schematic diagram of the updated damage distribution.

[0067] Figure 3 This is a structural diagram of a fatigue damage assessment and remaining life prediction system. Detailed Implementation

[0068] The present invention will be further described below with reference to specific embodiments.

[0069] The first embodiment of the present invention provides a method for assessing fatigue damage and predicting remaining life of structural components, comprising the following steps:

[0070] Step 1: Obtain the pulsating load signal at the critical point of the structural component. Each time the load returns to the zero position, it constitutes a cycle. Extract the stress amplitude in each stress cycle. Due to the randomness of the stress amplitude during the loading process, it is necessary to extrapolate the overall stress amplitude based on a certain sample size of stress amplitudes. Fit the overall stress amplitude using a probability density function. By using a certain number of samples to extrapolate the overall stress amplitude distribution, the stress amplitude characteristics of the structural component can be more completely expressed. Furthermore, the probability density function of the stress amplitude allows for real-time calculation of the probability and frequency of stress amplitude occurrences within any given time period. The damage within that time period can be obtained using the fatigue cumulative damage formula. Step 1 establishes the functional relationship between fatigue damage and time, allowing for real-time calculation of the fatigue damage D(t) of the structural component based on the stress probability density function.

[0071] The specific method is as follows:

[0072] First, the fatigue SN curve of the material is determined through fatigue tests on standard samples. This invention uses a power function to express the SN curve of the material:

[0073] s k N = c (1)

[0074] In the formula: s is the stress amplitude; N is the number of stress cycles that cause failure; k and c are material constants.

[0075] Based on the linear fatigue cumulative damage criterion, the fatigue cumulative damage under different stress amplitudes is calculated as follows:

[0076]

[0077] Where: D represents cumulative fatigue damage; n represents the stress amplitude level; i represents the i-th stress amplitude level; n i The stress amplitude s of level i i The number of stress cycles; N i For the specimen at stress amplitude s at level i i The number of cycles required for fatigue failure under the influence of the action.

[0078] From (1) and (2), we can obtain:

[0079]

[0080] In actual operation, the load process of construction machinery is generally an asymmetric pulsating cycle, meaning the minimum stress is 0, and there is no non-negative stress. The peak value of each cycle is the stress amplitude. Figure 1 As shown, ● represents the zero-return point and ○ represents the peak point.

[0081] Assuming the stress amplitude at time t is s(t), and the stress return-to-zero rate per unit time is f0, then the number of cycles at time t is f0t, and the probability density function of the stress amplitude is p(s). The probability that the stress amplitude lies within the interval [s, s+ds] is p(s)ds, where ds is a small increment of the stress amplitude. Therefore, the expected frequency n(s) of the stress amplitude s within time t is:

[0082] n(s)=f0tp(s)ds (4)

[0083] According to Miner's criterion, the damage increment caused by one stress cycle is 1 / N(s). The damage caused by a cycle with a stress amplitude of s within time t is:

[0084]

[0085] In the formula: N(s) is the number of cycles from when the stress amplitude is s until fatigue failure.

[0086] The formula for calculating the fatigue damage D(t) caused by all stress cycles within time t is:

[0087]

[0088] From equations (3) and (6), we can obtain:

[0089]

[0090] Step 2: The damage value at time t calculated in Step 1 using the fatigue cumulative damage formula is a constant. However, during actual use, structural components are affected by random factors such as operating loads, operations, and the environment. Therefore, the fatigue damage value of the same structural component after the same period of use may have multiple values, following a certain distribution characteristic. Based on Step 1, considering the randomness of the fatigue degradation process, it is treated as a stochastic process, and the Wiener stochastic process is used to model the fatigue damage process within this time period. Specifically, this time period is divided into r equal time segments. The fatigue damage within each time segment is calculated according to the method in Step 1, and the Wiener stochastic process is used to model the fatigue damage process within this segment, obtaining model parameters and the fatigue damage distribution at time t.

[0091] The specific method is as follows:

[0092] This invention uses the Wiener process to model the fatigue degradation process, and the Wiener process model is as follows:

[0093] D(t)=D(0)+μt+σB(t) (8)

[0094] In the formula: D(0) is the initial degradation state of the structural component, and this invention assumes that D(0) = 0; t is time; μ is the drift coefficient, which is used to characterize the degradation rate of the structural component; σ is the diffusion coefficient, which is used to characterize the randomness of the fatigue degradation process in time; B(t) is the standard Wiener process, which follows a normal distribution with a mean of 0 and a variance of t, denoted as N(0,t), and characterizes the randomness of the degradation of the structural component itself.

[0095] The fatigue damage D calculated at time t is obtained from step (1). t =D(t=t) i ), t i The value range of is from the 0th time to the rth time, i = 0, 1, 2, ..., r, denoted as From time 0 to time t r The historical degradation dataset. Let ΔD t =D t -D t-1 For time t i-1 By time t i The degradation increment, Δt i =t i -t i-1 Then, according to the properties of the Wiener process, ΔD t Follows the mean μΔt i The variance is σ 2 Δt i The normal distribution is denoted as ΔD. t ~N(μΔt) i ,σ 2 Δt i If the degradation increment ΔD is given, then the degradation increment is ΔD. t The probability density function is:

[0096]

[0097] Degradation increment ΔD at each time point t The joint probability density function is:

[0098]

[0099] The parameters are estimated using the maximum likelihood estimation method, and the following results are obtained:

[0100]

[0101] Based on the fatigue damage degradation process, the current state is estimated from the state at the previous moment, and a recursive equation for the damage relationship at each moment is established:

[0102] D t =D t-1 +μ t , where μt ~N(μ,σ 2 (12)

[0103] Where μ t Let t be the damage rate at time t.

[0104] The predicted fatigue damage distribution follows a normal distribution, denoted as:

[0105] D t |D t-1 ,μ t ~N(μ1,σ1) 2 )=N(μt,σ 2 t) (13)

[0106] Where μ1=μt, σ1 2 =σ 2 t, μ1, σ1 2 These represent the mean and variance of fatigue damage calculated based on the load information, respectively.

[0107] Step 3: Steps 1 and 2 are fatigue damage calculations obtained through theoretical models based on fatigue damage mechanisms and damage models. Since there are discrepancies between theoretical models and reality, relying solely on theoretical models may yield erroneous results. This invention proposes obtaining the current damage state of the structural component through detection methods to further confirm the fatigue damage process of the structural component.

[0108] The specific method is as follows:

[0109] The fatigue damage state at time t is detected using a nonlinear ultrasonic testing system, and the nonlinear ultrasonic coefficient is obtained. Based on the relationship between the nonlinear ultrasonic coefficient and fatigue damage, the fatigue damage distribution detected by nonlinear ultrasonic testing at time t is established.

[0110] The nonlinear ultrasonic coefficient β at time t was obtained by using a nonlinear ultrasonic testing system to test the structural component. t Based on the measurement accuracy of the nonlinear ultrasonic detection coefficient, it can be seen that the nonlinear ultrasonic coefficient at time t follows a mean of μ. βt The variance is σ βt 2 Normal distribution:

[0111] β t ~N(μβ) t ,σ βt 2 (14)

[0112] Since nonlinear ultrasonic testing does not directly measure damage values, it is necessary to convert the nonlinear ultrasonic coefficients into damage values. The damage value Z in nonlinear ultrasonic testing of structural components is thus determined. t With nonlinear ultrasonic coefficient βt The relationship can be described as follows:

[0113] Z t =g(β) t (15)

[0114] Wherein, g(β) t ) is the nonlinear ultrasonic coefficient β t The relationship between fatigue and damage is characterized.

[0115] At this point, the mean and variance of fatigue damage obtained from the nonlinear ultrasonic coefficients are respectively:

[0116] μ2=E(Z t )=E(g(β t (16)

[0117] σ2 2 =Var(Z) t )=Var(g(β t (17)

[0118] The fatigue damage detected by nonlinear ultrasound follows a normal distribution, denoted as Z. t |D t ~N(μ2,σ2) 2 )

[0119] Step 4: To reduce the errors introduced by using a single method to calculate the fatigue damage of structural components in Steps 2 and 3, this invention fuses the two fatigue damage results obtained in Steps 2 and 3. It fully utilizes the fatigue cumulative damage calculation results and the nonlinear ultrasonic testing results to obtain a fused fatigue damage distribution. Based on the principle of maximizing probability, the fused damage is determined as the current true damage distribution of the structural component. This improves the accuracy of fatigue damage calculation, reduces calculation errors, and facilitates the development of more refined maintenance / replacement strategies. The specific method is as follows:

[0120] Based on step (3), the fatigue damage prediction results of step (2) are updated according to the nonlinear ultrasonic detection results. The fatigue damage state D of the composite structural component at time t-1 is then determined. t-1 Damage rate μ t The current nonlinear ultrasonic damage detection result Z t Estimate the most likely fatigue damage state D. t ,Right now:

[0121] max P(D t |D t-1 Z t ,μ t (18)

[0122] According to conditional probability:

[0123]

[0124] Nonlinear ultrasonic damage value Z at time t t Only with the current damage state D t Related to the state and damage rate of the structural component at time t-1, the partial conditional probability in equation (19) can be transformed into: P(Z t |D t D t-1 ,μ t )=P(Z t |D t ), P(Z t |D t-1 ,μ t )=P(Z t ),but:

[0125]

[0126] Because P(Z) t If ) is an independent quantity, then equation (20) can be rewritten as:

[0127] P(D t |D t-1 Z t ,μ t )∝P(Z t |D t )P(D t |D t-1 ,μ t ) (twenty one)

[0128] Therefore, the final fatigue damage state of the structural component is a fusion of the fatigue damage calculation value based on the Wiener process and the fatigue damage detection value based on nonlinear ultrasonic testing, which is the product of two probabilities. Due to the influence of prediction error and measurement error, the prediction result in step (2) and the nonlinear ultrasonic testing result in step (3) at time t both follow a normal distribution, and the fused distribution is as follows: Figure 2 The dashed line shown is denoted as L. t This represents the distribution of the most likely outcome obtained by combining two different methods. According to the properties of the normal distribution, since P(D) in Formula 21... t |D t-1 Z t ,μ t Since the conditional law is proportional to the product of the two conditions, the overlapping portion also follows a normal distribution. The most likely fatigue damage occurs at the mean of this distribution. Therefore, the mean of this distribution can be used as an estimate of the most likely fatigue damage state at time t.

[0129]

[0130] Among them, the average fatigue damage after fusion Fatigue damage variance after fusion

[0131] Variance of fatigue damage after fusion By performing equivalent changes, we can obtain:

[0132]

[0133] Therefore, the prediction results are updated by using the fatigue damage detection results of nonlinear ultrasound. The updated result is a weighted fusion of the predicted value and the nonlinear ultrasound detection value. The overall variance of the updated result is smaller, the data accuracy is improved, and the result is more consistent with reality.

[0134] 5. Fatigue failure of structural components can lead to major safety accidents, causing casualties and huge economic losses. To improve the availability and safety of structural components, preventative maintenance or replacement is necessary to avoid failure during use. Therefore, this invention derives the remaining life distribution function of the structural component based on its fatigue damage threshold and current fatigue damage value, calculates the average remaining life of the component, and provides a 95% one-sided confidence lower limit for the remaining life. Maintenance or replacement plans can be formulated based on the average remaining life of the structural component, allowing for advance preparation of spare parts, personnel allocation, and maintenance or replacement strategies to minimize downtime losses and improve the safety and availability of the structural components.

[0135] If a structural component has not failed at time t, its remaining lifespan is predicted based on its current usage to support predictive maintenance and prevent significant economic losses due to component failure. The failure threshold L of the structural component is determined based on historical information and failure mechanisms. Structural component failure is defined as the service life of the component when fatigue damage first exceeds the failure threshold L. Let T be the service life of the structural component, which can be expressed as:

[0136] T=inf{t:D(t)≥L;t≥0} (24)

[0137] As can be seen from the properties of the Wiener process, the structural component lifetime T follows an inverse Gaussian distribution, and its probability density function is:

[0138]

[0139] The cumulative fatigue damage at time t obtained from step (4) is μ Σ The failure threshold L has not been exceeded. Due to the homogeneous Markov property of the Wiener process, the remaining life T of the structural component is... t for:

[0140] Tt =inf{t|D(t)≥L-μ Σ ;t≥0} (26)

[0141] Its probability density function also follows an inverse Gaussian distribution:

[0142]

[0143] Using the average degradation rate of structural components as a parameter for predicting the remaining life of structural components, the average remaining life of structural components can be obtained from equation (27):

[0144]

[0145] The 95% one-sided confidence lower limit for remaining lifetime is where α = 0.05, U 1-α It is the upper 1-α quantile of the standard normal distribution.

[0146] Based on the remaining life prediction results, maintenance plans are developed for structural components to ensure their safe operation and improve equipment utilization.

[0147] A second embodiment of the present invention provides a structural component fatigue damage assessment and remaining life prediction system, comprising a load acquisition module, a cycle counting module, a fatigue damage calculation module, a fatigue degradation modeling module, a nonlinear ultrasonic testing module, a nonlinear coefficient calculation module, a fatigue damage conversion module, a fatigue damage update module, a remaining life prediction module, and a maintenance decision module. Figure 3 As shown, the functions of each module are as follows:

[0148] Load acquisition module: Real-time monitoring of the stress state of structural components and data acquisition, obtaining the stress amplitude in each stress cycle of the structural components.

[0149] Cycle counting and distribution fitting module: Stress amplitude counting and distribution fitting, fitting the probability density function of stress amplitude based on stress amplitude, and obtaining the fatigue damage calculation formula based on stress amplitude and the probability density function of stress amplitude.

[0150] Fatigue damage calculation module: Calculates fatigue damage based on the Miner criterion, obtains the Wiener process model based on the fatigue degradation process of the structural component, and obtains the fatigue damage distribution.

[0151] Fatigue degradation modeling module: Performs degradation modeling on fatigue damage at each time step and evaluates degradation model parameters. Calculates fatigue damage of structural components at each time step according to fatigue damage calculation formula, and calculates parameters of the Wiener process model based on fatigue damage of structural components at each time step.

[0152] Nonlinear ultrasonic testing coefficient measurement module: Determines and acquires the nonlinear ultrasonic coefficient of the structural component.

[0153] Fatigue damage conversion module: Determines the fatigue damage of structural components based on the nonlinear ultrasonic coefficient, and establishes the fatigue damage distribution of nonlinear ultrasonic detection based on the relationship between the nonlinear ultrasonic coefficient and the nonlinear ultrasonic detection damage value.

[0154] Fatigue damage update module: Updates fatigue damage results based on nonlinear ultrasonic testing results. Based on the fatigue damage distribution calculation results and nonlinear ultrasonic testing damage values, obtains the corrected fatigue damage conditional probability based on the probability maximization principle. Merges the fatigue damage distribution with the nonlinear ultrasonic testing fatigue damage distribution based on the corrected fatigue damage conditional probability to obtain the corrected fatigue damage distribution. Calculates the current fatigue damage value of the structural component based on the corrected fatigue damage distribution.

[0155] Remaining life prediction module: It uses the fatigue degradation process to predict the remaining life. Based on the fatigue damage threshold and the current fatigue damage value of the structural component, it obtains the remaining life distribution function of the structural component and calculates the remaining life of the structural component according to the remaining life distribution function.

[0156] Maintenance Decision Module: Based on the remaining life prediction results, a maintenance / replacement plan for structural components is formulated in advance. This invention provides a method and system for assessing fatigue damage and predicting the remaining life of structural components, which offers the following advantages:

[0157] (1) Miner’s fatigue cumulative damage criterion can be replaced by a nonlinear fatigue cumulative damage criterion, such as the Carten-Dolan theory;

[0158] (2) The Wiener fatigue damage process can be replaced by nonlinear processes such as Gamma and inverse Gaussian, and the damage distribution function can be represented by a nonnormal distribution.

[0159] (3) The fatigue damage recursive prediction equation is linear and can be replaced by a nonlinear function. The nonlinear function can be transformed into a linear or approximately linear function, such as Taylor expansion at the mean point.

[0160] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for assessing fatigue damage and predicting remaining life of structural components, characterized in that: Includes the following steps: Step 1: Obtain the stress amplitude in each stress cycle of the structural component, fit the probability density function of the stress amplitude based on the stress amplitude, and obtain the fatigue damage calculation formula based on the stress amplitude and the probability density function of the stress amplitude. Step 2: Based on the fatigue degradation process of the structural component, obtain the Wiener process model, calculate the fatigue damage of the structural component at each time step according to the fatigue damage calculation formula, calculate the parameters of the Wiener process model based on the fatigue damage of the structural component at each time step, and obtain the fatigue damage distribution based on the parameters. Step 3: Obtain the nonlinear ultrasonic coefficient of the structural component, and establish the fatigue damage distribution of nonlinear ultrasonic testing based on the relationship between the nonlinear ultrasonic coefficient and the nonlinear ultrasonic detection damage value; Step 4: Based on the fatigue damage distribution calculation results and the nonlinear ultrasonic detection damage value, obtain the corrected fatigue damage conditional probability based on the probability maximization principle. Then, fuse the fatigue damage distribution and the nonlinear ultrasonic detection fatigue damage distribution according to the corrected fatigue damage conditional probability to obtain the corrected fatigue damage distribution. Finally, calculate the current fatigue damage value of the structural component based on the corrected fatigue damage distribution. Step 4 includes: Calculate the fatigue damage state at times t and t-1 based on the fatigue damage distribution. , Obtain the nonlinear ultrasonic damage value at time t. Based on the principle of probability maximization, the corrected fatigue damage conditional probability is obtained. The formula for calculating the corrected fatigue damage conditional probability is as follows: ; in, Let be the damage rate at time t; The fatigue damage distribution is obtained by fusing the fatigue damage distribution with the fatigue damage distribution detected by nonlinear ultrasonic testing based on the modified fatigue damage conditional probability. The formula for calculating the modified fatigue damage distribution is as follows: ; in, , ; The current fatigue damage value of the structural component is calculated based on the corrected fatigue damage distribution.

2. The method for assessing fatigue damage and predicting remaining life of structural components according to claim 1, characterized in that: Also includes: Step 5: Based on the fatigue damage threshold and the current fatigue damage value of the structural component, obtain the remaining life distribution function of the structural component, and calculate the remaining life of the structural component according to the remaining life distribution function.

3. The method for assessing fatigue damage and predicting remaining life of structural components according to claim 1 or 2, characterized in that: Step 1 includes: Obtain the stress amplitude of the structural component in each stress cycle. The probability density function of stress amplitude The fatigue damage calculation formula is obtained as follows: ; in, Let be the fatigue damage caused by all stress cycles at time t, and k and c be material constants. The stress return to zero rate per unit time. It is a tiny increment of the stress amplitude.

4. A method for assessing fatigue damage and predicting remaining life of structural components according to claim 1 or 2, characterized in that: Step 2 includes: Based on the fatigue degradation process of the structural components, a Wiener process model is obtained, and the calculation formula for the Wiener process model is as follows: ; in, The fatigue damage caused by all stress cycles at time t. This represents the initial degradation state of the structural component. The drift coefficient, Where is the diffusion coefficient. For standard Wiener procedures; Calculate the fatigue damage at time t using the fatigue damage calculation formula. , The value range is from the 0th time to the rth time, i = 0, 1, 2, ..., r; set up For time Time The increase in degradation, ; The parameters of the Wiener process model are calculated using the following formula: ; according to The fatigue damage distribution is obtained, and the formula for calculating the fatigue damage distribution is as follows: ; in, Let be the damage rate at time t. , .

5. A method for assessing fatigue damage and predicting remaining life of structural components according to claim 1 or 2, characterized in that: Step 3 includes: The nonlinear ultrasonic coefficient of the structural component at time t is obtained as follows: According to the nonlinear ultrasonic coefficient Calculation of damage values ​​from nonlinear ultrasonic testing , The calculation formula is as follows: ; in, To use nonlinear ultrasonic coefficients The relationship between fatigue and damage is characterized; Damage values ​​were detected using nonlinear ultrasonic testing. The fatigue damage distribution detected by nonlinear ultrasonic testing was established, and the calculation formula for the fatigue damage distribution detected by nonlinear ultrasonic testing is as follows: ; in, , .

6. The method for assessing fatigue damage and predicting remaining life of structural components according to claim 2, characterized in that: Step 5 includes: Based on the current fatigue damage value of the structural component and the failure threshold L of the structural component, the remaining life distribution function of the structural component is obtained. The formula for calculating the remaining life distribution function of the structural component is as follows: ； in, The drift coefficient, Let be the diffusion coefficient, and t be the time. ; The remaining life is obtained based on the remaining life distribution function of the structural component. The remaining life includes the average remaining life and the 95% one-sided confidence lower limit of the remaining life. The formula for calculating the average remaining life is as follows: ； The formula for calculating the 95% one-sided confidence lower limit of the remaining lifetime is as follows: 。 7. A system for assessing fatigue damage and predicting remaining life of structural components, characterized in that: Includes the following modules: Load acquisition module: used to acquire the stress amplitude of the structural component in each stress cycle; Cycle counting and distribution fitting module: used to fit the probability density function of stress amplitude based on stress amplitude, and obtain the fatigue damage calculation formula based on stress amplitude and the probability density function of stress amplitude; Fatigue damage calculation module: used to obtain the Wiener process model based on the fatigue degradation process of the structural component, and obtain the fatigue damage distribution; Fatigue degradation modeling module: used to calculate the fatigue damage of structural components at each time step according to the fatigue damage calculation formula, calculate the parameters of the Wiener process model based on the fatigue damage of structural components at each time step, and send the parameters to the fatigue damage calculation module; Nonlinear ultrasonic testing coefficient measurement module: used to determine the nonlinear ultrasonic coefficient of structural components and obtain the nonlinear ultrasonic coefficient of structural components; Fatigue damage conversion module: used to establish the fatigue damage distribution of nonlinear ultrasonic detection based on the relationship between the nonlinear ultrasonic coefficient and the nonlinear ultrasonic detection damage value; Fatigue damage update module: Based on the fatigue damage distribution calculation results and the nonlinear ultrasonic detection damage value, it obtains the corrected fatigue damage conditional probability based on the probability maximization principle, fuses the fatigue damage distribution and the nonlinear ultrasonic detection fatigue damage distribution according to the corrected fatigue damage conditional probability to obtain the corrected fatigue damage distribution, and calculates the current fatigue damage value of the structural component based on the corrected fatigue damage distribution. The fatigue damage update module includes: Calculate the fatigue damage state at times t and t-1 based on the fatigue damage distribution. , Obtain the nonlinear ultrasonic damage value at time t. Based on the principle of probability maximization, the corrected fatigue damage conditional probability is obtained. The formula for calculating the corrected fatigue damage conditional probability is as follows: ； in, Let be the damage rate at time t; The fatigue damage distribution is obtained by fusing the fatigue damage distribution with the fatigue damage distribution detected by nonlinear ultrasonic testing based on the modified fatigue damage conditional probability. The formula for calculating the modified fatigue damage distribution is as follows: ； in, , ; The current fatigue damage value of the structural component is calculated based on the corrected fatigue damage distribution.

8. The fatigue damage assessment and remaining life prediction system for structural components according to claim 7, characterized in that: Also includes: The remaining life prediction module is used to obtain the remaining life distribution function of the structural component based on the fatigue damage threshold and the current fatigue damage value, and to calculate the remaining life of the structural component based on the remaining life distribution function.

9. The fatigue damage assessment and remaining life prediction system for structural components according to claim 7, characterized in that: The cycle counting and distribution fitting module includes the following functions: Obtain the stress amplitude of the structural component in each stress cycle. The probability density function of stress amplitude The fatigue damage calculation formula is obtained as follows: ； in, Let be the fatigue damage caused by all stress cycles at time t, and k and c be material constants. The stress return to zero rate per unit time. It is a tiny increment of the stress amplitude.

Images