A reliability analysis method for aircraft chassis with multi-stage corrosion degradation

Abstract

The present invention proposes a reliability analysis method for an onboard trunk skin with multi-stage corrosion degradation, including: step one: establishment of corrosion degradation indicators and selection of random process models; specifically: establishment of corrosion degradation indicators; including data preprocessing and establishment of corrosion degradation indicators: selection of random process models; including random process models for corrosion degradation modeling, such as gamma process, inverse Gaussian process and Wiener process; step two: parameter estimation of a two-stage corrosion degradation model; specifically: normal stage model parameter estimation; defect stage model parameter estimation; stage change point determination; step three: full life cycle reliability calculation; specifically: full life cycle degradation process characterization; normal stage reliability calculation; defect stage reliability calculation; full life cycle reliability calculation; step four: reliability analysis of an onboard trunk skin based on two-stage corrosion degradation; specifically: sensitivity analysis of drift and diffusion coefficients; sensitivity analysis of stage change point parameters.

Description

Technical Field

[0001] The invention provides a reliability analysis method for an aircraft trunk with multi-stage corrosion degradation, relates to a reliability analysis technology based on multi-stage random process degradation modeling, and belongs to the field of reliability engineering. Background Art

[0002] Corrosion is a common phenomenon in industry and daily life. Its occurrence usually brings serious economic losses and safety hazards. Therefore, the research and intervention on corrosion deserves continuous deepening. Multi-stage corrosion degradation is a typical degradation characteristic of aircraft luggage products. The most common is a two-stage degradation process including a defect stage. That is, before the aircraft luggage fails due to corrosion degradation, it usually goes through a normal stage and a defect stage. In the normal stage, the degradation process of the product is slow. Once the aircraft luggage has defects and enters the defect stage, the degradation process will be accelerated. If the aircraft luggage is not intervened or the intervention is inappropriate, it will eventually fail. Conducting reliability analysis on aircraft luggage products with multi-stage corrosion degradation can fully promote their reliability and safety.

[0003] Currently, research on reliability analysis of multi-stage corrosion degradation is insufficient, particularly in terms of multi-stage degradation modeling and model-based reliability analysis. Establishing a suitable degradation model based on the degradation data characteristics of aircraft luggage is a fundamental issue. Limited by the quantity and quality of data, stochastic process-based degradation modeling is a more feasible approach. In multi-stage degradation, the determination of stage change points and the increasing number of stages present certain difficulties in degradation process modeling and reliability analysis. Using stochastic process theory to model the two-stage corrosion degradation process, deriving the reliability of aircraft luggage products throughout their lifecycle, and conducting intervention analysis on sensitive parameters can not only provide important guidance for intervention measures (such as maintenance) during product use, but also provide feedback to the design and production processes, further improving the inherent reliability of the product. Summary of the Invention

[0004] (1) Purpose of the present invention: This invention uses random process and reliability theory to carry out two-stage corrosion degradation process modeling, reliability index calculation and analysis work for a typical two-stage corrosion degradation product - aircraft trunk skin, so as to promote the improvement of its actual working health level.

[0005] (2) Technical solution:

[0006] The technical solution of the present invention is as follows: A reliability analysis method for an aircraft case with multi-stage corrosion degradation comprises the following steps:

[0007] Step 1: Establishment of corrosion degradation index and selection of random process model; specifically includes the following steps:

[0008] a) Establishment of corrosion degradation indicators; specifically including data preprocessing and establishment of corrosion degradation indicators:

[0009] First, data preprocessing is performed to obtain the original data of corrosion and degradation of the aircraft chassis, and abnormal values generated during the data recording process are eliminated by setting a continuity criterion.

[0010] The continuity criterion means that corrosion degradation will not suddenly increase sharply, that is, the difference between two adjacent degradation increments will not be too large. This limit is represented by ε. Different corrosion degradation data have different corresponding ε values, as shown in the following formula:

[0011]

[0012] In the formula, i, j, k represent different adjacent monitoring moments, X i ,X j ,X k They represent the corrosion degradation of the aircraft case corresponding to the three monitoring moments respectively.

[0013] The second step is to establish corrosion degradation indicators: The corrosion degradation process of aircraft luggage products generates various types of data. Corrosion degradation indicators are established from two perspectives: data characteristics and degradation mechanisms. Widely used indicators include corrosion depth, corrosion length, current, and insulation layer resistance. From the perspective of data characteristics, the aforementioned alternative indicators are selected based on characteristics that change significantly with the degradation process. From the perspective of degradation mechanisms, data types whose logic of data change is consistent with the corrosion degradation mechanism are selected (for example, for a volatile corrosion degradation process, the selected corrosion degradation indicator should generally also be volatile). The corrosion degradation indicator is determined by combining these two factors.

[0014] b) Model selection of random process: The random process models used for corrosion degradation modeling mainly include gamma process, inverse Gaussian process and Wiener process.

[0015] Gamma process: x(t)-x(0)~Ga(αt,β), where Ga(αt,β) represents the gamma distribution with shape parameter αt and scale parameter β;

[0016] Inverse Gaussian process: x(t)-x(0)~IG(μΛ(t),λ(Λ(t)) 2 ), where IG(μΛ(t),λ(Λ(t)) 2 ) represents the inverse Gaussian distribution, Λ(t) is a monotonically increasing function of time t;

[0017] Wiener process: x(t)-x(0)=μt+σB(t), where μ is the drift coefficient of the Wiener process, σ is the diffusion coefficient, and B(t) is the standard Brownian motion.

[0018] Among them, the gamma process and the inverse Gaussian process are both suitable for monotonic corrosion degradation modeling, that is, the time-related degradation increments are all positive; the Wiener process is suitable for fluctuating degradation processes, and the Wiener process has good analytical properties, which facilitates the derivation of reliability parameters; for different airborne luggage products and different corrosion degradation indicators, the degradation characteristics show different laws, and the appropriate random process model is selected according to the specific corrosion object and corrosion degradation indicator.

[0019] Step 2: Parameter estimation of the two-stage corrosion degradation model; specifically includes the following steps:

[0020] a) Normal stage model parameter estimation

[0021] The relationship between the degradation amount in the normal stage is: x1(t)=x1(0)+μ1t+σ1B(t), where μ1 and σ1 are the drift coefficient and diffusion coefficient of the onboard box skin in the normal stage, x1(0) is the corrosion degradation amount of the box skin at the initial moment, x1(t) is the corrosion degradation amount of the onboard box skin at the current moment, and B(t) is the standard Brownian motion.

[0022] Using the maximum likelihood estimation method, which is widely used for parameter estimation of degradation models, the log-likelihood function of the normal phase is established:

[0023]

[0024] Where i represents the i-th detection point, j represents the j-th sample onboard luggage, ΔX 1,j,i It represents the degradation change ΔX of the jth sample box at the i-th detection point 1,j,i =X 1,j,i -X 1,j,i-1 , Δt j,i Indicates the time difference Δt between adjacent detection points of the jth sample box j,i =t j,i -t j,i-1 .X 1,j,i For the jth box skin at t j,i The degradation at that moment.

[0025] Further, the partial derivatives of the drift coefficient μ1 and the diffusion coefficient σ1 of the airborne box skin in the normal stage are calculated respectively, and the expressions after partial derivative are set to zero. The parameter estimation formula is as follows:

[0026]

[0027] Where m is the number of similar aircraft-mounted cases, n is j is the number of corrosion degradation data of the j-th aircraft case.

[0028] b) Defect stage model parameter estimation

[0029] The difference between the defective stage and the normal stage is that the degradation rate and volatility will increase significantly, resulting in a significant decrease in the reliability of the onboard case. The degradation relationship of this stage is: x2(t) = x2(0) + μ2t + σ2B(t), where μ2 and σ2 are the drift coefficient and diffusion coefficient of the onboard case in the defective stage, x2(0) is the corrosion degradation of the system when it just enters the defective stage, and x2(t) is the corrosion degradation of the case at the current moment. Generally, μ2>μ1, σ2>σ1;

[0030] The parameter estimation in the defect phase is similar to that in the normal phase, except that the degradation data used does not include the normal phase and x2(0) is usually not 0. The estimated expressions for the drift coefficient and diffusion coefficient in the defect phase are:

[0031]

[0032] Where m is the number of similar aircraft-mounted cases, n is j is the number of corrosion degradation data of the jth box skin, X 2,j,i The jth onboard case at t j,i The degradation value at the moment ΔX 2,j,i It represents the degradation change ΔX of the jth sample box at the i-th detection point 2,j,i =X 2,j,i -X 2,j,i-1 , Δt j,i Indicates the time difference Δt between adjacent detection points of the jth sample box j,i =t j,i -t j,i-1 .X 2,j,i is the jth box skin at defective stage t j,i The degradation at that moment.

[0033] c) Determination of stage change points

[0034] The dividing point between the normal stage and the defect stage is the stage change point. There are two methods to determine the change point based on the corrosion degradation characteristics of the product and the difference in data volume: one is based on the degradation amount angle and the other is based on the randomness angle.

[0035] Degradation angle: When the degradation of the onboard box skin reaches a certain level ω0, it enters the defect stage; in the case of Wiener process degradation modeling, the stage change point t can be determined ω0 The analytical expression of the probability density function is as follows:

[0036]

[0037] Where L0 is the corrosion degradation failure threshold of the aircraft chassis.

[0038] From the perspective of randomness: the stage change point is a random variable Z, the distribution of which is independent of the degradation of the onboard luggage cover. The stage change point data in different samples are fitted with the Weibull distribution, and the scale parameter η and shape parameter γ of the Weibull distribution are estimated. The probability density function of Z is further obtained as follows:

[0039]

[0040] Step 3: Calculate the reliability of the entire life cycle; specifically, the following steps are included:

[0041] a) Characterization of the degradation process throughout the entire life cycle

[0042] For the two-stage degradation with stage change points determined from the perspective of randomness, the degradation process over the entire life cycle can be expressed as follows:

[0043]

[0044] Where z is the occurrence time of the stage change point, X1(t) is the degradation amount of the onboard case in the normal stage, and X2(t|z) is the degradation amount of the onboard case in the defective stage.

[0045] For the case where the product stage change point is determined from the perspective of degradation amount, the degradation process of the aircraft luggage skin throughout its life cycle can be characterized as follows:

[0046]

[0047] Where, is the moment when the degradation of the onboard box skin reaches ω0, that is, the two-stage change point.

[0048] b) Reliability calculation in normal stage

[0049] The normal stage is a necessary stage for two-stage corrosion degradation products. The reliability of this stage is calculated as follows:

[0050]

[0051] Where Φ(x) is the distribution function of the standard normal distribution, L0 is the corrosion degradation failure threshold of the aircraft chassis, and P(.) represents the probability corresponding to the situation in the brackets.

[0052] c) Defective stage reliability calculation

[0053] The defective stage is not a necessary stage for aircraft luggage, but products that have experienced the defective stage must have also experienced the normal stage. The reliability function of the defective stage is calculated as follows:

[0054]

[0055] In the formula, R2(t|z) represents the reliability of the onboard box skin at the defect stage t, X2(tz) represents the difference in the corrosion degradation of the onboard box skin from the two-stage change point z to the current time t, and x z It represents the corrosion degradation amount corresponding to the two-stage change point.

[0056] d) Full life cycle reliability calculation

[0057] The reliability of the aircraft luggage cover during its entire life cycle is determined by referring to the reliability functions of the normal stage and the defect stage. The calculation formula is as follows:

[0058]

[0059] In the formula, R(t) represents the reliability expression of the system in the entire life cycle, Represents the reliability expression of the system in the normal stage, F Z (z) represents the cumulative distribution function of the two-stage change point z, Indicates the probability that the two-stage change point z is greater than the current time t, Represents the reliability function of the system in the defect stage, f z (z) is the probability density function of the two-stage change point random variable Z of the onboard box skin.

[0060] Step 4: Reliability analysis of the airborne chassis based on two-stage corrosion degradation; specifically includes the following steps:

[0061] a) Sensitivity analysis of drift and diffusion coefficient

[0062] There are two sets of drift coefficients and diffusion coefficients in the two-stage corrosion degradation model. The parameter estimation result A0 (i.e., the estimation results of the drift coefficient and diffusion coefficient in the normal stage and the defect stage) is used as the benchmark, and three levels are set with A0 / 2 as the step size. For different levels of the same parameter, the reliability relationship of the system in the entire life cycle is calculated respectively. By plotting the relationship curve of the influence of each parameter change on the change of system reliability, the significance relationship of the influence of proportional changes of different types of parameters (with A0 / 2 as the step size) on the reliability of the aircraft luggage product is analyzed. That is, the degree of influence of different parameter changes on the system reliability is ranked, and the sensitive parameters are screened out accordingly. Generally, the one or two parameters with the highest significance relationship ranking are selected to guide the intervention direction of multiple parameters in the design, production and management links, that is, the most sensitive parameters are given priority for adaptive intervention.

[0063] b) Sensitivity analysis of stage change point parameters

[0064] The two-stage change points have a profound impact on the corrosion degradation process of aircraft luggage products. From the perspective of randomness, the sensitivity of the Weibull distribution parameters obeyed by the stage change points to the reliability of the product throughout its life cycle is analyzed. That is, for the shape parameter γ and the scale parameter η, the change step size is set respectively (for example, 1 / 2 of the current level is the step size). The different level values of the same parameter are substituted into formula (11) to obtain the expression of the system reliability during the life cycle at different parameter levels. Based on this, a reliability change curve with time at different parameter levels is drawn. The different parameter changes are further ranked according to the degree of influence on the system reliability. The factor with the most significant influence on the reliability of the two parameters, that is, the most sensitive factor, is screened out, thereby guiding the intervention direction of the shape parameter and scale parameter in the design, production and management links, that is, giving priority to the adaptive intervention of the most sensitive parameter.

[0065] (3) Advantages and Efficacy: The present invention provides a reliability analysis method for aircraft chassis with multi-stage corrosion degradation, which has the following advantages:

[0066] ① This invention studies the reliability modeling, analysis and evaluation methods of the multi-stage corrosion degradation process that is widely present in industrial products, and provides a method for modeling and analyzing multi-stage corrosion degradation failures;

[0067] ② This invention completes the characterization of the multi-stage degradation process based on considering multiple change point determination methods, which is more suitable for complex and multi-scenario scenarios in actual engineering;

[0068] ③ This invention establishes a full life cycle reliability based on multi-stage corrosion degradation for aircraft luggage products, and conducts reliability analysis work such as sensitivity analysis, which promotes the improvement of the reliability level of aircraft luggage products in design, production and management;

[0069] ④ The method of the present invention is scientific, has good processability, and has broad promotion and application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 It is a flow chart of the method of the present invention.

[0071] Figure 2 This is a diagram of the full life cycle reliability calculation results of an embodiment of the present invention.

[0072] Figure 3a-3b 2 is a graph showing the sensitivity of the drift coefficient according to an embodiment of the present invention.

[0073] Figure 4a-4b 2 is a diffusion coefficient sensitivity result diagram of an embodiment of the present invention.

[0074] Figure 5a-5b 2 is a diagram showing sensitivity results of change point parameters according to an embodiment of the present invention. DETAILED DESCRIPTION

[0075] In order to make the above-mentioned objects and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0076] The present invention provides a reliability analysis method for aircraft chassis with multi-stage corrosion degradation, which is specifically implemented as follows: Figure 1 As shown;

[0077] Step 1: Establishment of corrosion degradation index and selection of random process model; specifically divided into the following steps:

[0078] a) This application was conducted on the corrosion process of a certain aircraft chassis. The available degradation data included corrosion depth, corrosion length, and insulation resistance. Based on the data variation characteristics and degradation mechanism, the insulation resistance was determined as the corrosion degradation indicator. Furthermore, based on the continuity criterion and reference to the data variation during the corrosion degradation process, the ε value was set to 10 to eliminate some outliers.

[0079] b) Since the degradation increment of the insulation layer resistance value is not unique in positive or negative, that is, the degradation increment has fluctuations, which is consistent with the characteristics of the Wiener random process, the Wiener random process is used in this embodiment to carry out degradation modeling of the two-stage corrosion degradation process.

[0080] Step 2: Parameter estimation of the two-stage corrosion degradation model; specifically includes the following steps:

[0081] a) Screen the data of the slow degradation stage in the corrosion degradation index, use the maximum likelihood estimation method, and according to the formula and The drift coefficient μ1 and diffusion coefficient σ1 of the Wiener degradation model of the computer carrier skin product in the normal stage are shown in Table 1:

[0082] Table 1 Estimation results of degradation parameters in normal stage

[0083] <![CDATA[μ1]]> <![CDATA[σ1]]> 0.008 0.02

[0084] b) Screen the data of the slow degradation stage in the corrosion degradation index, use the maximum likelihood estimation method, and according to the formula and The drift coefficient μ2 and diffusion coefficient σ2 of the Wiener degradation model in the computer carrier skin defect stage are shown in Table 2:

[0085] Table 2 Degradation parameter estimation results of defect stage

[0086] <![CDATA[μ2]]> <![CDATA[σ2]]> 0.06 0.04

[0087] c) From a randomness perspective, the demarcation point between the normal stage and the defect stage in this embodiment was determined. The value corresponding to the stage change point in each sample was estimated from the resistance value change curve. MATLAB was further used to perform a Weibull distribution curve fit on the screened data to determine the shape parameter m and scale parameter η of the Weibull distribution. The specific values are shown in Table 3:

[0088] Table 3 Estimation results of distribution parameters of stage change points

[0089] m η 235 3.2

[0090] According to the formula Determine the two-stage dividing point of the onboard box, that is, the probability density function of the random variable Z is:

[0091] Step 3: Calculate the reliability of the entire life cycle; specifically, the following steps are included:

[0092] a) Substituting the parameter results obtained in the previous estimation into formula (7), the relationship between the corrosion degradation amount of the aircraft trunk cover in the whole life cycle and the time change is obtained as follows:

[0093]

[0094] b) According to the formula The reliability of the computer carrier at any time during the normal phase, such as Figure 2 shown.

[0095] c) According to the formula The reliability of the computer carrier skin at any time during the defect stage, such as Figure 2 shown.

[0096] d) According to the formula

[0097] The reliability of the computer carrier at any time during its life cycle, such as Figure 2 As shown. Analysis revealed that the results calculated in steps b) and c) coincide with those in step d), proving the rationality of the formula. On the other hand, observing the reliability curve of the aircraft luggage cover over its entire life cycle reveals that the product remains in the normal phase for a long time, with high reliability during this phase. However, once a defect occurs, the corrosion degradation process is accelerated, and the reliability of the aircraft luggage cover decreases rapidly, ultimately approaching zero. This suggests that intervention in the product's use is necessary to delay the occurrence of defects as much as possible, or to repair or replace the product as soon as possible after a defect occurs, in order to improve reliability and safety.

[0098] Step 4: Reliability analysis of the airborne chassis based on two-stage corrosion degradation; specifically includes the following steps:

[0099] a) Sensitivity analysis of drift coefficient and diffusion coefficient: First, sensitivity analysis of drift coefficient in normal stage and defect stage is carried out. Among them, the values of μ1 are 0.004, 0.008 and 0.012; the values of μ2 are 0.03, 0.06 and 0.09; the values of σ1 are 0.01, 0.02 and 0.03; the values of σ2 are 0.02, 0.04 and 0.06; according to the formula

[0100] The reliability of the airborne case skin over its entire life cycle is calculated under different drift coefficient levels in normal and defective states. The results are as follows: Figure 3a-3b The reliability of the entire life cycle of the onboard luggage under different diffusion coefficients is further calculated, as shown in Figure 4a-4b As shown. By observing Figure 3a-3b and Figure 4a-4b It can be concluded that the changes in the drift coefficient and diffusion coefficient in the normal stage have a more significant impact on the reliability of the onboard box skin, because the system is in the normal stage most of the time; on the other hand, this also inspires us to make the drift coefficient and diffusion coefficient in the normal stage as low as possible through the design and production process, so as to achieve the purpose of improving inherent reliability and operational reliability.

[0101] c) Based on the estimated results of Weibull distribution parameters, the three levels of scale parameter are set to 118, 235 and 352, and the three levels of shape parameter are set to 1.6, 3.2 and 4.8. When the value of one parameter changes, the estimated results of the other parameter levels are maintained; according to the formula Calculate the reliability under different parameter levels. The influence of scale parameters on product reliability is more significant. Therefore, in the design of the product, Figure 5a-5b As shown. By observing Figure 5a-5b It can be concluded that the design, production and use processes should focus on the growth of scale parameters.

[0102] In summary, the present invention relates to a reliability analysis method for aircraft luggage with multi-stage corrosion degradation, which relates to a reliability analysis technology based on multi-stage random process degradation modeling. The specific steps of the method are: 1. Establishment of corrosion degradation indicators and selection of random process models; 2. Estimation of parameters of two-stage corrosion degradation model; 3. Calculation of full life cycle reliability; 4. Reliability analysis of aircraft luggage based on two-stage corrosion degradation. The present invention is applicable to the field of multi-stage degradation modeling and reliability analysis of aircraft luggage, and is of great significance for accurately evaluating the reliability of aircraft luggage, analyzing the influence of degradation model parameters on system reliability, and thus improving the reliability and health level of aircraft luggage in the design, production and use links; it can also be further extended to the degradation modeling and reliability analysis of other two-stage corrosion degradation products in industrial systems.

Claims

1. A reliability analysis method for aircraft chassis with multi-stage corrosion degradation, characterized in that: The following steps are involved: Step 1: Establishment of corrosion degradation indicators and selection of random process models. Specifically, the establishment of corrosion degradation indicators includes data preprocessing and establishment of corrosion degradation indicators, and random process model selection. Random process models used for corrosion degradation modeling include gamma process, inverse Gaussian process, and Wiener process. Step 2: Parameter estimation of the two-stage corrosion degradation model; specifically: normal stage model parameter estimation; defect stage model parameter estimation; stage change point determination; Step 3: Full life cycle reliability calculation; specifically: full life cycle degradation process characterization; normal stage reliability calculation; defect stage reliability calculation; full life cycle reliability calculation; Step 4: Reliability analysis of the airborne tank based on two-stage corrosion degradation; specifically: sensitivity analysis of drift and diffusion coefficients; sensitivity analysis of stage change point parameters; In step 3, the normal phase reliability is calculated as: The normal stage is a necessary stage for two-stage corrosion degradation products. The reliability of this stage is calculated as follows: Where Φ(x) is the distribution function of the standard normal distribution, L0 is the corrosion degradation failure threshold of the aircraft trunk skin, P(.) represents the probability corresponding to the situation in the brackets; X(t) is the degradation process of the entire life cycle; X1(t) is the degradation amount of the aircraft trunk skin in the normal stage; Z is a random variable; t is the current moment; μ1 and σ1 are the drift coefficient and diffusion coefficient of the aircraft trunk skin in the normal stage, respectively; In step 3, the defect stage reliability is calculated as: The reliability function of the defect stage is calculated as follows: In the formula, R2(t|z) represents the reliability of the onboard box skin at the defect stage t, X2(tz) represents the difference in the corrosion degradation of the onboard box skin from the two-stage change point z to the current time t, and x z represents the corrosion degradation amount corresponding to the change point of the two stages; μ2 and σ2 are the drift coefficient and diffusion coefficient of the airborne box skin defect stage respectively; In step 3, the life cycle reliability is calculated as: The reliability of the aircraft luggage cover during its entire life cycle is determined by referring to the reliability functions of the normal stage and the defect stage. The calculation formula is as follows: In the formula, R(t) represents the reliability expression of the system in the entire life cycle, Represents the reliability expression of the system in the normal stage, F Z (z) represents the cumulative distribution function of the two-stage change point z, Indicates the probability that the two-stage change point z is greater than the current time t, Represents the reliability function of the system in the defect stage, f z (z) is the probability density function of the two-stage change point random variable Z of the onboard box skin.

2. The reliability analysis method for aircraft chassis with multi-stage corrosion degradation according to claim 1 is characterized by: In step 1, the corrosion degradation index is established as follows: first, data preprocessing is performed to obtain the original data of corrosion degradation of the aircraft chassis, and abnormal values generated during the data recording process are eliminated by setting a continuity criterion; The continuity criterion means that corrosion degradation will not suddenly increase sharply, that is, the difference between two adjacent degradation increments will not be too large. This limit is represented by ε. Different corrosion degradation data have different corresponding ε values, as shown in the following formula: In the formula, i, j, k represent different adjacent monitoring moments, X i ,X j ,X k They represent the corrosion degradation of the aircraft case corresponding to the three monitoring moments respectively; The second step is to establish corrosion degradation indicators: The corrosion degradation process of aircraft luggage products will generate various types of data. Corrosion degradation indicators are established from two perspectives: data characteristics and degradation mechanisms. The data characteristics are selected from the characteristics of the degradation process data changes, and the degradation mechanism is selected from the logic of data changes and the data type that is suitable for the corrosion degradation mechanism. The corrosion degradation indicators are determined by combining the two. Model selection for random processes is: Gamma process: x(t)-x(0)~Ga(αt,β), where Ga(αt,β) represents the gamma distribution with shape parameter α, time t, and scale parameter β; Inverse Gaussian process: x(t)-x(0)~IG(μΛ(t),λ(Λ(t)) 2 ), where IG(μΛ(t),λ(Λ(t)) 2 ) represents the inverse Gaussian distribution, Λ(t) is a monotonically increasing function of time t; Wiener process: x(t)-x(0)=μt+σB(t), where μ is the drift coefficient of the Wiener process, σ is the diffusion coefficient, and B(t) is the standard Brownian motion; Among them, the gamma process and the inverse Gaussian process are both suitable for monotonic corrosion degradation modeling, that is, the time-related degradation increments are all positive; the Wiener process is suitable for fluctuating degradation processes.

3. The reliability analysis method for aircraft chassis with multi-stage corrosion degradation according to claim 2 is characterized by: In step 2, the normal stage model parameters are estimated as: The degradation relationship in the normal stage is: x1(t) = x1(0) + μ1t + σ1B(t), μ1 and σ1 are the drift coefficient and diffusion coefficient of the airborne tank skin in the normal stage, x1(0) is the corrosion degradation of the tank skin at the initial moment, x1(t) is the corrosion degradation of the airborne tank skin at the current moment, and B(t) is the standard Brownian motion. Using the maximum likelihood estimation method, the log-likelihood function of the normal phase is established: Where i represents the i-th detection point, j represents the j-th sample onboard luggage, ΔX 1,j,i It represents the degradation change ΔX of the jth sample box at the i-th detection point 1,j,i =X 1,j,i -X 1,j,i-1 , Δt j,i Indicates the time difference Δt between adjacent detection points of the jth sample box j,i =t j,i -t j,i-1 ;X 1,j,i For the jth box skin at t j,i The amount of degradation at a moment; Further, the partial derivatives of the drift coefficient μ1 and the diffusion coefficient σ1 of the airborne box skin in the normal stage are calculated respectively, and the expressions after partial derivative are set to zero. The parameter estimation formula is as follows: Where m is the number of similar aircraft-mounted cases, n is j is the number of corrosion degradation data of the j-th aircraft case.

4. The reliability analysis method for aircraft chassis with multi-stage corrosion degradation according to claim 3 is characterized by: In step 2, the defect stage model parameters are estimated as: The degradation relationship of this stage is: x2(t)=x2(0)+μ2t+σ2B(t), μ2 and σ2 are the drift coefficient and diffusion coefficient of the airborne tank skin defect stage, x2(0) is the corrosion degradation of the system when it just enters the defect stage, and x2(t) is the corrosion degradation of the tank skin at the current moment. So: μ2>μ1,σ2>σ1; The difference between the parameter estimation of the defective stage and the normal stage is that the degradation data used does not include the normal stage situation and x2(0) is not 0; The estimated expressions of the drift coefficient and diffusion coefficient in the defect phase are obtained: Where m is the number of similar aircraft-mounted cases, n is j is the number of corrosion degradation data of the jth box skin, X 2,j,i The jth onboard case at t j,i The degradation value at the moment; ΔX 2,j,i It represents the degradation change ΔX of the jth sample box at the i-th detection point 2,j,i =X 2,j,i -X 2,j,i-1 , Δt j,i Indicates the time difference Δt between adjacent detection points of the jth sample box j,i =t j,i -t j,i-1 ;X 2,j,i is the jth box skin at defective stage t j,i The degradation at that moment.

5. The reliability analysis method for aircraft chassis with multi-stage corrosion degradation according to claim 4 is characterized by: In step 2, the stage change point is: The dividing point between the normal stage and the defect stage is the stage change point. There are two methods to determine the change point based on the corrosion degradation characteristics of the product and the difference in data volume: one is based on the degradation amount angle and the other is based on the randomness angle. Degradation angle: When the degradation of the onboard box reaches a certain level ω0, it enters the defect stage; in the case of Wiener process degradation modeling, the stage change point is determined The analytical expression of the probability density function is as follows: Where L0 is the corrosion degradation failure threshold of the aircraft case; From the perspective of randomness: the stage change point is a random variable Z, the distribution of which is independent of the degradation of the onboard luggage cover. The stage change point data in different samples are fitted with the Weibull distribution, and the scale parameter η and shape parameter γ of the Weibull distribution are estimated. The probability density function of Z is further obtained as follows:

6. The reliability analysis method for aircraft chassis with multi-stage corrosion degradation according to claim 5 is characterized by: In step three, the full life cycle degradation process is characterized as follows: For the two-stage degradation with stage change points determined from the perspective of randomness, the full life cycle degradation process is expressed as follows: Where z is the occurrence time of the stage change point, X1(t) is the degradation amount of the onboard case in the normal stage, and X2(t|z) is the degradation amount of the onboard case in the defective stage. For the case where the product stage change point is determined from the perspective of degradation amount, the degradation process of the aircraft luggage skin throughout its life cycle is characterized as follows: Where, is the moment when the degradation of the onboard box skin reaches ω0, that is, the two-stage change point.

7. The reliability analysis method for aircraft chassis with multi-stage corrosion degradation according to claim 1 is characterized by: In step 4, the sensitivity analysis of drift and diffusion coefficient is: There are two sets of drift coefficients and diffusion coefficients in the two-stage corrosion degradation model. The parameter estimation result A0 is used as the benchmark, A0 / 2 is used as the step size, and three level values are set. Different level values of the same parameter are substituted into formula (11) to calculate the reliability relationship of the system in the whole life cycle. By drawing the relationship curve of the influence of each parameter change on the system reliability change, the significance relationship of the influence of different types of proportional changes on the reliability of the airborne luggage product is analyzed, that is, the degree of influence of different parameter changes on the system reliability is ranked, and the sensitive parameters are screened out accordingly. The one or two parameters with the highest significance ranking are selected to guide the intervention direction of multiple parameters in the design, production and management links, that is, the most sensitive parameters are given priority for adaptive intervention. The sensitivity analysis of the stage change point parameters is: For the shape parameter γ and scale parameter η, the change step size is set respectively. Different level values of the same parameter are substituted into formula (11) to obtain the expression of the system reliability over the entire life cycle at different parameter levels. Based on this, a reliability change curve with time at different parameter levels is drawn. The influence of different parameter changes on system reliability is further sorted, and the factor with the most significant influence on reliability among the two parameters, that is, the most sensitive factor, is screened out, thereby guiding the intervention direction of shape parameters and scale parameters in the design, production and management links, that is, giving priority to adaptive intervention on the most sensitive parameters.

Images