Time-varying reliability analysis method and system for functionally graded material cylinder
Through random process discretization and univariate dimensionality reduction, the problem of insufficient computational complexity and accuracy in the time-varying reliability analysis of cylinders of functional gradient materials is solved, and efficient and accurate time-varying reliability evaluation is achieved, suitable for nonlinearity and multimodal function functions.
Patent Information
- Application Number
- CN202411156079.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-22
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-08-22
AI Technical Summary
The existing time-varying reliability analysis methods have insufficient computational complexity and accuracy, especially in the cylindrical structure of functional gradient materials. Traditional methods are difficult to effectively deal with nonlinear and multimodal functional functions, resulting in high computational costs and no results. precise.
Using random process discretization and univariate dimensionality reduction methods, the discretization and dimensionality reduction processing are performed by defining the random variables and stochastic processes of the cylinder of the functional gradient material, and the statistical moment and probability density function of the instantaneous functional function are calculated to obtain the failure probability curve, realize time-varying reliability analysis.
It significantly improves the computational efficiency and accuracy, reduces the computational cost, and can accurately evaluate the time-varying reliability of cylinders of functional gradient materials, and is suitable for nonlinear and multimodal functional functions.
Smart Images

Figure CN119004835A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of time-varying reliability, and in particular to a time-varying reliability analysis method and system for a functional gradient material cylinder. Background Art
[0002] Cylindrical structures are widely used in engineering structures. It is of great significance to study the stability of cylindrical structures under various loads. Cylindrical structures are often used as supporting structures of buildings, such as bridges, buildings and gymnasiums. They can not only withstand huge pressures, but also provide good stability. Cylindrical storage containers, such as oil tanks, water towers and gas tanks, are ideal for storing liquids and gases because of their uniform structure and can effectively disperse internal and external pressures. The main bodies of rockets and spacecrafts are mostly cylindrical in design, which can reduce air resistance, increase flight stability and fuel efficiency. Cylindrical structures occupy an important position in modern engineering and technology due to their excellent physical properties and wide application prospects. On the other hand, due to the increasingly high requirements for structural performance in extreme environments and high-speed industries, cylindrical structures made of functionally gradient materials have received widespread attention in recent years. In order to ensure the safety and stability of functionally gradient cylindrical structures in practical applications, reliability analysis of the structure is an indispensable link in the research and development and manufacturing process.
[0003] The purpose of reliability analysis is to calculate the probability that a structure will achieve its intended function while taking into account various uncertainties such as material properties, manufacturing, loading conditions, and environment. Traditional reliability analysis methods assume that uncertainties do not take into account time and can only evaluate instantaneous reliability. However, the performance of actual engineering structures is also affected by time-varying or dynamic sources of uncertainty, such as time-varying loads or material degradation, and ignoring time parameters may lead to inaccurate reliability assessments. Time-varying reliability analysis is a reliability assessment method based on the entire life cycle or a specific time interval. Evaluating the probability within the time interval of interest requires repeating the instantaneous reliability assessment at each time node. In actual engineering, the limit state is generally difficult to evaluate, and considering the time, the limit state response will be a complex time-varying random variable or random process. The implementation of time-varying reliability analysis requires a large number of evaluations of the time-varying limit state function, which will lead to a huge computational burden. Therefore, it is a challenge to achieve high efficiency and accuracy in time-varying reliability analysis.
[0004] For time-varying reliability problems, existing calculation methods can be roughly divided into three categories: the first is the heterogeneity method, which calculates the failure probability that varies with time by integrating the heterogeneity rate (the ratio of the probability of the limit state event to the time), which can more accurately reflect the reliability change of the system on the time scale, but it needs to determine the form of the performance function and has a very high computational complexity. The second is the extreme value method, which obtains the distribution parameters of the extreme value by counting the number of times the system response extreme value exceeds the limit state in a given time interval, and transforms the time-varying reliability analysis problem into a traditional time-invariant (static) reliability analysis problem. It requires running a large number of simulation operations and has a huge computational cost. The third is the response agent-based method, which uses a proxy model instead of the real time-varying limit state function to improve efficiency. Despite the progress made in the past few decades, the above methods still have some shortcomings and limitations. For problems with nonlinear and multimodal performance functions, the heterogeneity method may obtain inaccurate results. Due to model assumptions, the empirical model used to calculate the failure probability may be inaccurate in some cases. The sample-based extreme value method may encounter bottlenecks in practical applications due to the large amount of calculation, while the agent-based extreme value method is still in the development stage in terms of ensuring the credibility of the model. Therefore, a time-varying reliability analysis method and system for functionally gradient material cylinders are proposed. Summary of the invention
[0005] The purpose of the present invention is to provide a time-varying reliability analysis method and system for a functional gradient material cylinder, which combines random process discretization and single variable dimension reduction method to calculate the failure probability, and has high efficiency and accuracy.
[0006] To achieve the above object, the present invention provides the following solutions:
[0007] A time-varying reliability analysis method for a functionally gradient material cylinder, comprising:
[0008] Define a random variable, a random process, and a time domain of a functionally gradient material cylinder, wherein the random variable is a material parameter of the functionally gradient material cylinder, the random process is a load on the functionally gradient material cylinder that varies with time, and the time domain is a preset time interval;
[0009] Discretizing the random process and the time domain to obtain a number of time nodes, and an instantaneous functional function and an independent random variable at each time node;
[0010] Performing dimensionality reduction processing on the instantaneous functional function based on a single variable dimensionality reduction method, and calculating the statistical moment of the instantaneous functional function;
[0011] Calculating the probability density function of the instantaneous functional function based on the statistical moment, and then calculating the instantaneous failure probability;
[0012] The instantaneous failure probabilities of all time nodes are combined to obtain the failure probability curve, and the time-varying reliability analysis of the functional gradient material cylinder is completed.
[0013] Optionally, the random variable is a material parameter of the functionally gradient material cylinder, the random process is a load on the functionally gradient material cylinder that varies with time, and the time domain is a preset time interval.
[0014] Optionally, after the instantaneous function is reduced in dimension based on a single variable dimensionality reduction method, a number of one-dimensional integrals are obtained, and the reduced instantaneous function is obtained by solving the sum of the number of one-dimensional integrals. The reduced instantaneous function is:
[0015]
[0016] Among them, g(X) is the instantaneous function after dimensionality reduction, μ is the mean of the random variable, g(μ1,...,x i ,...μ n ) represents the random variable x i is a univariate function of , where n is the number of random variables.
[0017] Optionally, the method for calculating the statistical moment of the instantaneous functional function is:
[0018]
[0019] Among them, l represents the order of the statistical moment, represents the calculation of permutations and combinations, and E[·] represents the moment of ·.
[0020] Optionally, the method for calculating the probability density function of the instantaneous functional function based on the statistical moment is:
[0021] max H[g(X)]=-∫ R p[g(X)]lnp[g(X)]dX
[0022] st∫ R [g(X)] j p[g(X)]dX=μ j j=0,1,…,m
[0023] Among them, max H[g(X)] is the maximum entropy of the instantaneous performance function, p[g(X)] is the probability density function of the instantaneous performance function, μ j is the j-th order statistical moment, and m is the highest order of the statistical moment taken.
[0024] Optionally, the method for calculating the instantaneous failure probability is:
[0025]
[0026] Among them, P f is the instantaneous failure probability, and Z is an independent random variable.
[0027] To further achieve the above-mentioned purpose, the present invention also provides a time-varying reliability analysis system for a functionally gradient material cylinder, comprising: an information definition module, an information processing module, and an information analysis module;
[0028] The information definition module is used to define the dynamic information of the functional gradient material cylinder;
[0029] The information processing module is used to discretize and reduce the dimension of the dynamic information, and calculate the instantaneous failure probability of each time node of the functional gradient material cylinder;
[0030] The information analysis module is used to combine the instantaneous failure probabilities of all time nodes to obtain a failure probability curve.
[0031] Optionally, the dynamic information includes random variables, random processes, and a time domain, wherein the random variables are material parameters of the functional gradient material cylinder, the random process is the load on the functional gradient material cylinder that changes with time, and the time domain is a preset time interval.
[0032] Optionally, the information processing module includes: a discretization unit, a dimension reduction unit, and a calculation unit;
[0033] The discretization unit is used to discretize the random process and the time domain to obtain a number of time nodes, and the instantaneous function and independent random variables of each time node;
[0034] The dimension reduction unit is used to perform dimension reduction processing on the instantaneous functional function to obtain a one-dimensional function;
[0035] The calculation unit is used to calculate the statistical moment of the instantaneous functional function based on the one-dimensional function; and calculate the probability density function of the instantaneous functional function based on the statistical moment, so as to calculate the instantaneous failure probability.
[0036] The beneficial effects of the present invention are:
[0037] The present invention aims at the problem of large amount of calculation and high degree of function nonlinearity in time-varying reliability analysis of functional gradient cylindrical structure, and proposes a time-varying reliability analysis method based on dimensionality reduction. By discretizing the random process, the time-varying function function is converted into multiple instantaneous function functions at different times. Since the random process requires additional equivalent random variables to represent, in order to avoid the large amount of calculation cost required for reliability solution, the probability density curve of all instantaneous functions is calculated using the single variable dimensionality reduction method, and the failure probability at each moment is obtained. Not only is the efficiency greatly improved compared with the traditional method, but it can also maintain a high calculation accuracy for the problems of nonlinear and multimodal function functions. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0039] Figure 1 A flow chart of a time-varying reliability analysis method for a functionally gradient material cylinder according to an embodiment of the present invention;
[0040] Figure 2 This is a schematic diagram of a functionally gradient cylindrical structure according to an embodiment of the present invention;
[0041] Figure 3 It is a comparison diagram of failure probability curves of the method according to an embodiment of the present invention and MCS. DETAILED DESCRIPTION
[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0043] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.
[0044] This embodiment provides a time-varying reliability analysis method for a functionally gradient material cylinder. Figure 1 As shown, including:
[0045] Define random variables, random processes, and time domains of a functionally gradient material cylinder, wherein the random variables are material parameters of the functionally gradient material cylinder, the random processes are time-varying loads on the functionally gradient material cylinder, and the time domain is a preset time interval;
[0046] Discretize the random process and time domain to obtain several time nodes, as well as the instantaneous function and independent random variables of each time node;
[0047] The instantaneous function is reduced in dimension based on the single variable dimension reduction method, and the statistical moment of the instantaneous function is calculated;
[0048] Calculate the probability density function of the instantaneous functional function based on the statistical moment, and then calculate the instantaneous failure probability;
[0049] The instantaneous failure probabilities of all time nodes are combined to obtain the failure probability curve and complete the time-varying reliability analysis of the functionally gradient material cylinder.
[0050] Specifically, by discretizing the random process, the time-varying function is converted into multiple instantaneous function functions at different times. Since the random process requires additional equivalent random variables to represent, in order to avoid the large amount of computational cost required for reliability solution, the probability density curve of all instantaneous functions is calculated using the single variable dimensionality reduction method, and the failure probability at each moment is obtained. Not only is the efficiency greatly improved compared with traditional methods, but it can also maintain high computational accuracy for nonlinear and multimodal function problems.
[0051] The following is Figure 2 Taking the functionally gradient material cylindrical structure shown in the figure as an example, a time-varying reliability analysis method of a functionally gradient material cylinder proposed in this embodiment is described in detail, including the following steps:
[0052] Step 1: Set random variables;
[0053] The material parameters of the functionally gradient material are taken as random variables X{x i}, including the elastic modulus, Poisson's ratio, density, etc. of the material. The form of material parameters is:
[0054] E=E1V1+E2V2
[0055] ν=ν1V1+ν2V2
[0056] ρ=ρ1V1+ρ2V2
[0057] Where E, ν, ρ, and V are Young's modulus, Poisson's ratio, density, and volume fraction, respectively. Subscripts 1 and 2 represent two materials, respectively.
[0058] Step 2: Set boundary conditions, including random process and time domain;
[0059] The time-varying load is regarded as the random process Y(t), and the distribution form and autocorrelation function in the time domain T are given.
[0060] The functionally graded material cylindrical structure is subjected to a lateral force Y(t) that varies with time. The distribution of random parameters is shown in Table 1.
[0061] Table 1
[0062]
[0063] The autocorrelation function of T(t) is expressed as:
[0064] ρ(t1,t2)=exp[-(t2-t1) 2 ]
[0065] The random process is expressed as:
[0066]
[0067] Where p is the number of dominant eigenvalues and is less than N t ,λ h and Φ h are the hth eigenvalue and eigenvector of the covariance matrix, respectively, and the (k,l)th element is σ y (t k )σ y (t l )σ y (t k ,t l ). Y (t) is given by [σ Y (t)σ Y (t1)σ Y (t,t1),...,σ Y (t)σ Y (t n )σ Y (t,t n )] T Compute the time-varying covariance vector of Z. h is the hth independent standard normal random variable.
[0068] Step 3: Discretize the random process and time domain to obtain several time nodes, as well as the instantaneous function and independent random variables of each time node;
[0069] In the time interval [t0,t e ] uniform selection within N t time instant, N t The number of N determines the accuracy of the discretization of the random process.t The larger the value, the more accurately the dependencies between different instantaneous function functions are captured.
[0070] According to the random process obtained by discretization, the dynamic calculation function g(x,Y(t),t) of the functional gradient material cylinder is transformed into g={g(x,Z,t),t k ∈[t0,t e ],k=1,...,N t}, where t k The instantaneous performance function at g(x,Z,t) is related to the performance functions at other moments. The dependencies between different instantaneous performance functions can be implicitly captured in the discretization of Y(t). Although Y(t) is represented by a set of independent random variables Z at each time instant, the coefficients of Z are extracted from the covariance matrix, whose elements contain the relevant information about Y(t) at different time instants. The coefficients of Z remain correlated between discrete g(x,Z,t).
[0071] The dynamic response of the functional gradient material cylinder, i.e., the functional function, is calculated based on solving the dynamic control equation. In this embodiment, the dynamic control equation is solved using the differential orthogonal method and the Newmark method. The dynamic equation can be expressed as:
[0072]
[0073] Where σ, ρ and u are stress, density and displacement respectively, and the relationship between strain and displacement is expressed as:
[0074]
[0075] Where ε is the strain, and the physical equation can be expressed as
[0076]
[0077] Substituting the physical equation and strain-displacement relationship into the control equation, we get
[0078]
[0079] The initial boundary conditions are expressed as follows:
[0080]
[0081] The stress-based performance function is defined as
[0082]
[0083] Where r represents the radial axis in the cylindrical coordinate system, and the displacement u at any coordinate is obtained by the differential orthogonal method and the Newmark method.
[0084] Step 4: Perform dimensionality reduction processing on the instantaneous functional function based on the single variable dimensionality reduction method, and calculate the statistical moment of the instantaneous functional function;
[0085] The single variable dimensionality reduction method is used to calculate the discrete g(x, Z, t). The main idea of the single variable dimensionality reduction method is to transform the multidimensional integral into several one-dimensional integrals, approximate the original function by solving the sum of these one-dimensional integrals, and obtain the instantaneous functional function after dimensionality reduction.
[0086] The approximation of the instantaneous function g(X) after dimensionality reduction takes the following form:
[0087]
[0088] where μ is the mean of the random variable. g(μ1,...,x i ,...μ n ) represents the random variable x i The univariate function of the instantaneous function can be expressed as follows after reducing the dimension of the instantaneous function:
[0089]
[0090] Where l represents the order of the statistical moment, represents the random variable x i The integral formula is calculated by Gaussian integral, so the integral expression for moment is:
[0091]
[0092] Step 5: Calculate the probability density function of the instantaneous functional function based on the statistical moment;
[0093] The maximum entropy method is used to calculate the probability density function (PDF) of the functional function. According to the information theory principle, entropy is defined as the expression of the probability density function:
[0094] H=-∫p[g(X)]lnp[g(X)]dX
[0095] Under different constraints, the expression of maximizing entropy can get the distribution expression of minimizing prior information. For continuous variables, the expression of maximum entropy method is:
[0096] max H[g(X)]=-∫ R p[g(X)]lnp[g(X)]dX
[0097] st∫ R [g(X)] j p[g(X)]dX=μ j j=0,1,…,m
[0098] This optimization problem can be solved by using Lagrange function and Lagrange multiplier, namely:
[0099]
[0100] Among them, there are (N+1) Lagrange multipliers λ=[λ0,λ1,...,λ n ],φ n (x) takes the form of φ n [g(X)]=[g(X)] n p[g(X)] is the function of the maximum entropy probability density required, μ j It corresponds to the j-th order statistical moment, m is the highest order of the statistical moment, usually 4th order. The constraint problem can be solved iteratively using the standard Newton method.
[0101] Step 6: Calculate the instantaneous failure probability;
[0102] Substitute the performance function probability density function (PDF) into the following formula:
[0103]
[0104] The failure probability of the instantaneous limit state function can be obtained.
[0105] Step 7: Combine the instantaneous failure probabilities of all time nodes, obtain the failure probability curve, and complete the time-varying reliability analysis of the functionally gradient material cylinder.
[0106] To further optimize the above technical solution, the present invention also provides a time-varying reliability analysis system for a functionally gradient material cylinder, including: an information definition module, an information processing module, and an information analysis module;
[0107] An information definition module is used to define the dynamic information of a functionally gradient material cylinder;
[0108] An information processing module, used for discretizing and reducing the dimension of the dynamic information, and calculating the instantaneous failure probability of each time node of the functional gradient material cylinder;
[0109] The information analysis module is used to combine the instantaneous failure probabilities of all time nodes and obtain the failure probability curve.
[0110] Specifically, the dynamic information includes random variables, random processes, and time domains, wherein the random variables are material parameters of the functional gradient material cylinder, the random process is the load on the functional gradient material cylinder that changes with time, and the time domain is a preset time interval.
[0111] Specifically, the information processing module includes: a discretization unit, a dimension reduction unit, and a calculation unit;
[0112] The discretization unit is used to discretize the random process and the time domain to obtain a number of time nodes, and the instantaneous function and independent random variables of each time node;
[0113] The dimension reduction unit is used to perform dimension reduction processing on the instantaneous functional function to obtain a one-dimensional function; the calculation unit is used to calculate the statistical moment of the instantaneous functional function based on the one-dimensional function; and calculate the probability density function of the instantaneous functional function based on the statistical moment, and then calculate the instantaneous failure probability.
[0114] Comparing the method proposed in this embodiment with the results of MCS, Table 2 shows the error of failure probability. Figure 3 The comparison of failure probability curves is given. It can be seen that the maximum error with the MCS result is 3.76%, indicating that this method has high accuracy in solving the time-varying reliability analysis problem of functional gradient material cylinder dynamics. In terms of computational efficiency, the number of times MCS calls the functional function is 21×10 6 times, while the number of function calls of this method is only 21×313=6573 times, which greatly reduces the calculation cost.
[0115] Table 2
[0116]
[0117] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.
Claims
1. A time-varying reliability analysis method for a functionally gradient material cylinder, characterized in that: include: Define a random variable, a random process, and a time domain of a functionally gradient material cylinder, wherein the random variable is a material parameter of the functionally gradient material cylinder, the random process is a load on the functionally gradient material cylinder that varies with time, and the time domain is a preset time interval; Discretizing the random process and the time domain to obtain a number of time nodes, and an instantaneous functional function and an independent random variable at each time node; Performing dimensionality reduction processing on the instantaneous functional function based on a single variable dimensionality reduction method, and calculating the statistical moment of the instantaneous functional function; Calculating the probability density function of the instantaneous functional function based on the statistical moment, and then calculating the instantaneous failure probability; The instantaneous failure probabilities of all time nodes are combined to obtain the failure probability curve, and the time-varying reliability analysis of the functional gradient material cylinder is completed.
2. The time-varying reliability analysis method of a functionally gradient material cylinder according to claim 1, characterized in that: After the instantaneous function is reduced in dimension based on the single variable dimensionality reduction method, a number of one-dimensional integrals are obtained, and the reduced instantaneous function is obtained by solving the sum of the several one-dimensional integrals. The reduced instantaneous function is: Among them, g(X) is the instantaneous function after dimensionality reduction, μ is the mean of the random variable, g(μ1,...,x i ,...μ n ) represents the random variable x i is a univariate function of , where n is the number of random variables.
3. The time-varying reliability analysis method of a functionally gradient material cylinder according to claim 2, characterized in that: The method for calculating the statistical moment of the instantaneous functional function is: Among them, l represents the order of the statistical moment, represents the calculation of permutations and combinations, and E[·] represents the moment of ·.
4. The time-varying reliability analysis method of a functionally gradient material cylinder according to claim 3, characterized in that: The method for calculating the probability density function of the instantaneous functional function based on the statistical moment is: max H[g(X)]=-∫ R p[g(X)]ln p[g(X)]dX s.t.∫ R [g(X)] j p[g(X)]dX=μ j j=0,1,…,m Among them, max H[g(X)] is the maximum entropy of the instantaneous performance function, p[g(X)] is the probability density function of the instantaneous performance function, μ j is the j-th order statistical moment, and m is the highest order of the statistical moment taken.
5. The time-varying reliability analysis method of a functionally gradient material cylinder according to claim 4, characterized in that: The method for calculating the instantaneous failure probability is: Among them, P f is the instantaneous failure probability, and Z is an independent random variable.
6. A time-varying reliability analysis system for a functionally gradient material cylinder, used to implement the time-varying reliability analysis method for a functionally gradient material cylinder as claimed in any one of claims 1 to 5, characterized in that: include: Information definition module, information processing module, information analysis module; The information definition module is used to define the dynamic information of the functional gradient material cylinder; The information processing module is used to discretize and reduce the dimension of the dynamic information, and calculate the instantaneous failure probability of each time node of the functional gradient material cylinder; The information analysis module is used to combine the instantaneous failure probabilities of all time nodes to obtain a failure probability curve.
7. The time-varying reliability analysis system for a functionally gradient material cylinder according to claim 6, characterized in that: The dynamic information includes random variables, random processes, and a time domain, wherein the random variables are material parameters of the functionally gradient material cylinder, the random process is a load on the functionally gradient material cylinder that changes with time, and the time domain is a preset time interval.
8. The time-varying reliability analysis system for a functionally gradient material cylinder according to claim 7, characterized in that: The information processing module includes: a discretization unit, a dimension reduction unit, and a calculation unit; The discretization unit is used to discretize the random process and the time domain to obtain a number of time nodes, and the instantaneous function and independent random variables of each time node; The dimension reduction unit is used to perform dimension reduction processing on the instantaneous functional function to obtain a one-dimensional function; The calculation unit is used to calculate the statistical moment of the instantaneous functional function based on the one-dimensional function; and calculate the probability density function of the instantaneous functional function based on the statistical moment, so as to calculate the instantaneous failure probability.
