Probabilistic Engineering Mechanics最新文献_第2页

Confidence-based design optimization using multivariate kernel density estimation under insufficient input data 在输入数据不足的情况下，利用多变量核密度估计进行基于置信度的设计优化

IF 3 3区工程技术 Q2 ENGINEERING, MECHANICAL

Probabilistic Engineering Mechanics

Pub Date : 2024-10-01 DOI: 10.1016/j.probengmech.2024.103702

Yongsu Jung , Minjik Kim , Hyunkyoo Cho , Weifei Hu , Ikjin Lee

The uncertainty quantification of the input statistical model in reliability-based design optimization (RBDO) has been widely investigated for accurate reliability analysis, and it could be estimated through its characteristics, cumulative experiences, and available data. However, uncertainty quantification of random variables in existing RBDO studies has exploited parametric distributions quantifying the uncertainty through the Bayes' theorem. In addition, a correlation between random variables is often underestimated due to a lack of knowledge and difficulty to describe the high-dimensional correlation. Hence, it has been a challenge to properly quantify input statistical model and its uncertainty. Therefore, a multivariate kernel density estimation (KDE) is employed to perform data-driven confidence-based design optimization (CBDO) for effective quantification of input model uncertainty. Any assumption on input distribution is not necessary since it is established only with the given input data. Moreover, the input model uncertainty due to insufficient data is quantified using bootstrapping and optimal adaptive bandwidth matrices through the Bayes’ theorem using cross-validation error. Consequently, the proposed CBDO with given input data is capable of finding a conservative optimum of RBDO accounting for both aleatory uncertainty of random variables and epistemic uncertainty induced by a limited number of input data through the multivariate KDE.

为进行精确的可靠性分析，基于可靠性的设计优化（RBDO）中输入统计模型的不确定性量化已被广泛研究，可通过其特征、累积经验和可用数据对其进行估计。然而，在现有的 RBDO 研究中，随机变量的不确定性量化都是利用参数分布，通过贝叶斯定理量化不确定性。此外，由于缺乏知识和难以描述高维相关性，随机变量之间的相关性往往被低估。因此，如何正确量化输入统计模型及其不确定性一直是个难题。因此，我们采用了多元核密度估计（KDE）来执行数据驱动的基于置信度的设计优化（CBDO），以有效量化输入模型的不确定性。由于只需根据给定的输入数据即可建立输入分布，因此无需对输入分布进行任何假设。此外，数据不足导致的输入模型不确定性可通过贝叶斯定理使用交叉验证误差进行引导和最优自适应带宽矩阵量化。因此，所提出的具有给定输入数据的 CBDO 能够找到 RBDO 的保守最优值，既考虑到随机变量的可知不确定性，也考虑到通过多元 KDE 由有限的输入数据引起的可知不确定性。

{"title":"Confidence-based design optimization using multivariate kernel density estimation under insufficient input data","authors":"Yongsu Jung , Minjik Kim , Hyunkyoo Cho , Weifei Hu , Ikjin Lee","doi":"10.1016/j.probengmech.2024.103702","DOIUrl":"10.1016/j.probengmech.2024.103702","url":null,"abstract":"<div><div>The uncertainty quantification of the input statistical model in reliability-based design optimization (RBDO) has been widely investigated for accurate reliability analysis, and it could be estimated through its characteristics, cumulative experiences, and available data. However, uncertainty quantification of random variables in existing RBDO studies has exploited parametric distributions quantifying the uncertainty through the Bayes' theorem. In addition, a correlation between random variables is often underestimated due to a lack of knowledge and difficulty to describe the high-dimensional correlation. Hence, it has been a challenge to properly quantify input statistical model and its uncertainty. Therefore, a multivariate kernel density estimation (KDE) is employed to perform data-driven confidence-based design optimization (CBDO) for effective quantification of input model uncertainty. Any assumption on input distribution is not necessary since it is established only with the given input data. Moreover, the input model uncertainty due to insufficient data is quantified using bootstrapping and optimal adaptive bandwidth matrices through the Bayes’ theorem using cross-validation error. Consequently, the proposed CBDO with given input data is capable of finding a conservative optimum of RBDO accounting for both aleatory uncertainty of random variables and epistemic uncertainty induced by a limited number of input data through the multivariate KDE.</div></div>","PeriodicalId":54583,"journal":{"name":"Probabilistic Engineering Mechanics","volume":"78 ","pages":"Article 103702"},"PeriodicalIF":3.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142437726","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A data-driven maximum entropy method for probability uncertainty analysis based on the B-spline theory 基于 B-样条理论的概率不确定性分析数据驱动最大熵方法

IF 3 3区工程技术 Q2 ENGINEERING, MECHANICAL

Probabilistic Engineering Mechanics

Pub Date : 2024-10-01 DOI: 10.1016/j.probengmech.2024.103688

Gang Li , Yiyuan Wang , Wanxin He , Changting Zhong , Yixuan Wang

The probability density function (PDF) is quite important for structural reliability analysis; thus, accurate PDF modeling methods draw increasing attention. This paper proposes a novel metaheuristic data-driven paradigm of the maximum entropy method (MEM) based on the B-spline function theory. Firstly, a B-spline proxy of the MEM PDF is proposed for probability uncertainty analysis. We derive the parameter calculation formulation and calculate the undetermined parameters via the raw data of structural responses. Then, to determine the knots of the B-spline functions, we propose a novel data-driven approach with the aid of a powerful metaheuristic algorithm and the response data information. Different from the traditional MEM, the proposed method is a complete data-driven solution approach and does not involve the statistical moment calculation and the nonlinear equations composed of statistical moments. Combining the advantages of the B-spline theory and the MEM, the proposed method can reconstruct the response PDF with a complex shape, such as the PDF with multiple peaks or heavy tails. For verification, two numerical examples and one engineering example are analyzed, and compared with some classical PDF modeling methods. The results show that the proposed method is superior to the compared methods in terms of computational accuracy, when the same sample data is used.

概率密度函数（PDF）对于结构可靠性分析相当重要，因此精确的 PDF 建模方法越来越受到关注。本文基于 B-样条函数理论，提出了一种新颖的元启发式数据驱动最大熵方法（MEM）范式。首先，提出了用于概率不确定性分析的 MEM PDF 的 B 样条代理。我们推导出参数计算公式，并通过结构响应的原始数据计算未确定参数。然后，为了确定 B-样条函数的节点，我们提出了一种新颖的数据驱动方法，借助强大的元启发式算法和响应数据信息。与传统的 MEM 方法不同，所提出的方法是一种完整的数据驱动求解方法，不涉及统计矩计算和由统计矩组成的非线性方程。结合 B-样条理论和 MEM 的优点，所提出的方法可以重构具有复杂形状的响应 PDF，如具有多个峰值或重尾的 PDF。为进行验证，分析了两个数值实例和一个工程实例，并与一些经典的 PDF 建模方法进行了比较。结果表明，在使用相同样本数据的情况下，所提出的方法在计算精度方面优于其他方法。

{"title":"A data-driven maximum entropy method for probability uncertainty analysis based on the B-spline theory","authors":"Gang Li , Yiyuan Wang , Wanxin He , Changting Zhong , Yixuan Wang","doi":"10.1016/j.probengmech.2024.103688","DOIUrl":"10.1016/j.probengmech.2024.103688","url":null,"abstract":"<div><div>The probability density function (PDF) is quite important for structural reliability analysis; thus, accurate PDF modeling methods draw increasing attention. This paper proposes a novel metaheuristic data-driven paradigm of the maximum entropy method (MEM) based on the B-spline function theory. Firstly, a B-spline proxy of the MEM PDF is proposed for probability uncertainty analysis. We derive the parameter calculation formulation and calculate the undetermined parameters via the raw data of structural responses. Then, to determine the knots of the B-spline functions, we propose a novel data-driven approach with the aid of a powerful metaheuristic algorithm and the response data information. Different from the traditional MEM, the proposed method is a complete data-driven solution approach and does not involve the statistical moment calculation and the nonlinear equations composed of statistical moments. Combining the advantages of the B-spline theory and the MEM, the proposed method can reconstruct the response PDF with a complex shape, such as the PDF with multiple peaks or heavy tails. For verification, two numerical examples and one engineering example are analyzed, and compared with some classical PDF modeling methods. The results show that the proposed method is superior to the compared methods in terms of computational accuracy, when the same sample data is used.</div></div>","PeriodicalId":54583,"journal":{"name":"Probabilistic Engineering Mechanics","volume":"78 ","pages":"Article 103688"},"PeriodicalIF":3.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142421054","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Closed-form expressions for eigenvalue and eigenvectors of stochastic symmetric matrices using the probability transformation method 使用概率变换法的随机对称矩阵特征值和特征向量闭式表达式

IF 3 3区工程技术 Q2 ENGINEERING, MECHANICAL

Probabilistic Engineering Mechanics

Pub Date : 2024-10-01 DOI: 10.1016/j.probengmech.2024.103706

Rossella Laudani, Giovanni Falsone

This work shows the use of the Probability Transformation Method (PTM) for deriving a closed-form probability density function (PDF) of the eigenpair of stochastic real-valued symmetric matrices. In particular, the PTM allows the direct evaluation of the eigenpair PDF starting from the joint PDF (JPDF) of the system’s uncertainties. The impact of the linear stochastic systems’ randomness in the natural frequencies and mode shape is investigated through some numerical applications. Even if the structural samples investigated are intentionally simple, that aspect is only linked to the authors’ use of the Mathematica software that, in some ways, limits the resolution for high dimensional problems. From a theoretical perspective, though, this is not a restriction, and the problem’s dimension has no impact on the method’s accuracy. The obtained analytical results compared with Monte Carlo simulations have confirmed the goodness of the proposed stochastic procedure.

这项工作展示了如何利用概率变换法（PTM）推导出随机实值对称矩阵特征对的闭式概率密度函数（PDF）。特别是，PTM 可以从系统不确定性的联合概率密度函数（JPDF）出发，直接评估特征对的概率密度函数。通过一些数值应用研究了线性随机系统的随机性对固有频率和模态振型的影响。尽管所研究的结构样本有意简单化，但这只是与作者使用的 Mathematica 软件有关，该软件在某些方面限制了高维问题的分辨率。不过，从理论角度来看，这并不是一个限制，问题的维度对方法的准确性没有影响。所获得的分析结果与蒙特卡罗模拟结果相比较，证实了所提出的随机程序的优越性。

引用次数: 0

Quantitative property of MF-discrepancy and efficient point-selection strategy for the nonlinear stochastic response analysis of structures with random parameters 随机参数结构非线性随机响应分析的中频差异定量特性和高效选点策略

IF 3 3区工程技术 Q2 ENGINEERING, MECHANICAL

Probabilistic Engineering Mechanics

Pub Date : 2024-10-01 DOI: 10.1016/j.probengmech.2024.103708

Jian-Bing Chen , Xin Huang , Jie Li

The response analysis of high-dimensional and strongly nonlinear systems with random parameters remains a significant challenge in stochastic computational mechanics. To address this challenge, some methods based on the high-efficacy point sets have been developed, in which efficient global-point-set methods represented by low-discrepancy are of paramount importance in generating representative point sets. Several discrepancies including the extended F-discrepancy (EF-discrepancy) and the generalized F-discrepancy (GF-discrepancy) have been introduced to assess the uniformity and the efficacy of a representative point set. In such context, a maximal marginal EF-discrepancy (MF-discrepancy), which is an extended form of the GF-discrepancy, is proposed in this paper and then the properties of the MF-discrepancy are studied in detail. The probability distribution of the MF-discrepancy is derived, including a rigorous proof for random point sets and a model based on an assumption for some generic point sets. A generalized Koksma-Hlawka inequality is established accordingly to govern the worst error estimate. The lowest bound of the MF-discrepancy is given, and two intuitive quantitative indices are proposed to measure the goodness of the MF-discrepancy. Based on the lowest bound, an enhanced point-selection strategy with a unified theoretical framework for minimizing the MF-discrepancy is proposed. In this framework, locally minimizing the MF-discrepancy yields the two-step point-selection method, and a new point-selection strategy is proposed based on the global minimization of the MF-discrepancy, which is verified to be efficient and robust, especially in high-dimensional cases. Several numerical examples, including a 2-story shear frame, a 10-story shear frame, and a 10-story reinforced concrete frame structure modeled by the finite element method, are studied, verifying the efficiency and the robustness of the proposed point-selection strategy.

对具有随机参数的高维强非线性系统进行响应分析，仍然是随机计算力学领域的一项重大挑战。为了应对这一挑战，人们开发了一些基于高效点集的方法，其中以低差异为代表的高效全局点集方法对生成代表性点集至关重要。为了评估代表性点集的均匀性和有效性，引入了几种差异，包括扩展 F-差异（EF-差异）和广义 F-差异（GF-差异）。在此背景下，本文提出了最大边际 EF-差异（MF-差异），它是 GF-差异的扩展形式，然后详细研究了 MF-差异的特性。本文推导了 MF-差异的概率分布，包括随机点集的严格证明和基于一些通用点集假设的模型。据此建立了广义的 Koksma-Hlawka 不等式，以控制最差误差估计。给出了中频差异的最低界限，并提出了两个直观的量化指标来衡量中频差异的好坏。在最低界限的基础上，提出了一种具有统一理论框架的增强型选点策略，以最小化中频差异。在这一框架中，局部最小化中频差异产生了两步选点法，并基于中频差异的全局最小化提出了一种新的选点策略，该策略被证实是高效和稳健的，尤其是在高维情况下。研究了几个数值实例，包括一个 2 层剪力框架、一个 10 层剪力框架和一个用有限元法建模的 10 层钢筋混凝土框架结构，验证了所提出的选点策略的效率和鲁棒性。

{"title":"Quantitative property of MF-discrepancy and efficient point-selection strategy for the nonlinear stochastic response analysis of structures with random parameters","authors":"Jian-Bing Chen , Xin Huang , Jie Li","doi":"10.1016/j.probengmech.2024.103708","DOIUrl":"10.1016/j.probengmech.2024.103708","url":null,"abstract":"<div><div>The response analysis of high-dimensional and strongly nonlinear systems with random parameters remains a significant challenge in stochastic computational mechanics. To address this challenge, some methods based on the high-efficacy point sets have been developed, in which efficient global-point-set methods represented by low-discrepancy are of paramount importance in generating representative point sets. Several discrepancies including the extended F-discrepancy (EF-discrepancy) and the generalized F-discrepancy (GF-discrepancy) have been introduced to assess the uniformity and the efficacy of a representative point set. In such context, a maximal marginal EF-discrepancy (MF-discrepancy), which is an extended form of the GF-discrepancy, is proposed in this paper and then the properties of the MF-discrepancy are studied in detail. The probability distribution of the MF-discrepancy is derived, including a rigorous proof for random point sets and a model based on an assumption for some generic point sets. A generalized Koksma-Hlawka inequality is established accordingly to govern the worst error estimate. The lowest bound of the MF-discrepancy is given, and two intuitive quantitative indices are proposed to measure the goodness of the MF-discrepancy. Based on the lowest bound, an enhanced point-selection strategy with a unified theoretical framework for minimizing the MF-discrepancy is proposed. In this framework, locally minimizing the MF-discrepancy yields the two-step point-selection method, and a new point-selection strategy is proposed based on the global minimization of the MF-discrepancy, which is verified to be efficient and robust, especially in high-dimensional cases. Several numerical examples, including a 2-story shear frame, a 10-story shear frame, and a 10-story reinforced concrete frame structure modeled by the finite element method, are studied, verifying the efficiency and the robustness of the proposed point-selection strategy.</div></div>","PeriodicalId":54583,"journal":{"name":"Probabilistic Engineering Mechanics","volume":"78 ","pages":"Article 103708"},"PeriodicalIF":3.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142663357","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A perspective on conditional spectrum-based determination of response statistics of nonlinear systems 基于条件谱确定非线性系统响应统计的视角

IF 3 3区工程技术 Q2 ENGINEERING, MECHANICAL

Probabilistic Engineering Mechanics

Pub Date : 2024-10-01 DOI: 10.1016/j.probengmech.2024.103704

Beatrice Pomaro , Pol D. Spanos

This work focuses on determining the stochastic response properties, in the frequency domain, of a general class of nonlinear systems with polynomial nonlinearities. Specifically, the results are presented in terms of the stationary power spectral densities of the system's displacement and velocity. This is pursued by revisiting the conditional power spectrum concept, with the assumption that the response process is both ergodic and pseudo-harmonic and characterized by an amplitude, and a phase. A theoretical elucidation of an existing formula for the conditional spectrum is attempted. In particular, this concept is interpreted in conjunction with the time averaging approximation made in the definition of the stationary probability density function of a response amplitude quantity, associated with the original nonlinear system. It is shown that a proper definition of the stationary probability density of the response amplitude, along with a reasonable treatment of the distribution over the frequency domain of the amplitude contribution, lead to an improved approximation of the stationary response power spectral density. The treatment involves the averaging of a population of surrogate spectral densities of stationary random responses conforming with the system responses associated with individual values of the amplitudes of the responses. The semi-analytical results have been quite favourably juxtaposed with a large suite of à propos Monte Carlo simulations, both in terms of the shape and of the range of the involved germane frequencies, even for strongly nonlinear systems.

这项研究的重点是确定一类具有多项式非线性特性的通用非线性系统在频域上的随机响应特性。具体来说，研究结果是以系统位移和速度的静态功率谱密度来表示的。为此，我们重新审视了条件功率谱的概念，假设响应过程既是遍历的，又是伪谐波的，并以振幅和相位为特征。我们尝试从理论上阐明现有的条件功率谱公式。特别是，在定义与原始非线性系统相关的响应振幅量的静态概率密度函数时，结合时间平均近似对这一概念进行了解释。结果表明，正确定义响应振幅的静态概率密度，并合理处理振幅贡献的频域分布，可改进静态响应功率谱密度的近似值。这种处理方法包括对与响应振幅的各个值相关的系统响应相一致的大量静态随机响应的替代频谱密度进行平均。半分析结果与大量的蒙特卡罗模拟结果进行了比较，在涉及的相关频率的形状和范围方面，甚至在强非线性系统方面，结果都相当不错。

{"title":"A perspective on conditional spectrum-based determination of response statistics of nonlinear systems","authors":"Beatrice Pomaro , Pol D. Spanos","doi":"10.1016/j.probengmech.2024.103704","DOIUrl":"10.1016/j.probengmech.2024.103704","url":null,"abstract":"<div><div>This work focuses on determining the stochastic response properties, in the frequency domain, of a general class of nonlinear systems with polynomial nonlinearities. Specifically, the results are presented in terms of the stationary power spectral densities of the system's displacement and velocity. This is pursued by revisiting the conditional power spectrum concept, with the assumption that the response process is both ergodic and pseudo-harmonic and characterized by an amplitude, and a phase. A theoretical elucidation of an existing formula for the conditional spectrum is attempted. In particular, this concept is interpreted in conjunction with the time averaging approximation made in the definition of the stationary probability density function of a response amplitude quantity, associated with the original nonlinear system. It is shown that a proper definition of the stationary probability density of the response amplitude, along with a reasonable treatment of the distribution over the frequency domain of the amplitude contribution, lead to an improved approximation of the stationary response power spectral density. The treatment involves the averaging of a population of surrogate spectral densities of stationary random responses conforming with the system responses associated with individual values of the amplitudes of the responses. The semi-analytical results have been quite favourably juxtaposed with a large suite of à propos Monte Carlo simulations, both in terms of the shape and of the range of the involved germane frequencies, even for strongly nonlinear systems.</div></div>","PeriodicalId":54583,"journal":{"name":"Probabilistic Engineering Mechanics","volume":"78 ","pages":"Article 103704"},"PeriodicalIF":3.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142552656","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A physics-informed neural network enhanced importance sampling (PINN-IS) for data-free reliability analysis 用于无数据可靠性分析的物理信息神经网络增强型重要性采样（PINN-IS）

IF 3 3区工程技术 Q2 ENGINEERING, MECHANICAL

Probabilistic Engineering Mechanics

Pub Date : 2024-10-01 DOI: 10.1016/j.probengmech.2024.103701

Atin Roy , Tanmoy Chatterjee , Sondipon Adhikari

Reliability analysis of highly sensitive structures is crucial to prevent catastrophic failures and ensure safety. Therefore, these safety-critical systems are to be designed for extremely rare failure events. Accurate statistical quantification of these events associated with a very low probability of occurrence requires millions of evaluations of the limit state function (LSF) involving computationally expensive numerical simulations. Variance reduction techniques like importance sampling (IS) reduce such repetitions to a few thousand. The use of a data-centric metamodel can further cut it down to a few hundred. In data-centric metamodeling approaches, the actual complex numerical analysis is performed at a few points to train the metamodel for approximating the structural response. On the other hand, a physics-informed neural network (PINN) can predict the structural response based on the governing differential equation describing the physics of the problem, without a single evaluation of the complex numerical solver, i.e., data-free. However, the existing PINN models for reliability analysis have been effective only in estimating a large range of failure probabilities (10⁻¹∼10⁻³). To address this issue, the present study develops a PINN-based data-free reliability analysis for low failure probabilities (<10⁻⁵). In doing so, a two-stage PINN integrated with IS (PINN-IS) is proposed. The first stage is employed to approximate the most probable failure point (MPP) appropriately while the second stage enhances the accuracy of LSF predictions at the IS population centred on the approximated MPP. The effectiveness of the proposed approach is numerically illustrated by three structural reliability analysis examples.

高敏感结构的可靠性分析对于防止灾难性故障和确保安全至关重要。因此，这些安全关键型系统必须针对极其罕见的故障事件进行设计。要对这些发生概率极低的事件进行精确的统计量化，需要对极限状态函数（LSF）进行数百万次评估，其中涉及计算成本高昂的数值模拟。重要度采样（IS）等降低方差技术可将此类重复计算减少到数千次。使用以数据为中心的元模型可将重复次数进一步减少到几百次。在以数据为中心的元模型方法中，实际的复杂数值分析是在几个点上进行的，以训练元模型来逼近结构响应。另一方面，物理信息神经网络（PINN）可以根据描述问题物理特性的控制微分方程预测结构响应，而无需对复杂的数值求解器进行一次评估，即无需数据。然而，现有的用于可靠性分析的 PINN 模型只能有效估计较大范围的失效概率（10-1∼10-3）。针对这一问题，本研究开发了一种基于 PINN 的无数据可靠性分析方法，适用于低故障概率（10-5）。为此，我们提出了一种与 IS 集成的两阶段 PINN（PINN-IS）。第一阶段用于适当近似最可能故障点 (MPP)，第二阶段则提高以近似 MPP 为中心的 IS 群的 LSF 预测精度。通过三个结构可靠性分析实例对所提方法的有效性进行了数值说明。

{"title":"A physics-informed neural network enhanced importance sampling (PINN-IS) for data-free reliability analysis","authors":"Atin Roy , Tanmoy Chatterjee , Sondipon Adhikari","doi":"10.1016/j.probengmech.2024.103701","DOIUrl":"10.1016/j.probengmech.2024.103701","url":null,"abstract":"<div><div>Reliability analysis of highly sensitive structures is crucial to prevent catastrophic failures and ensure safety. Therefore, these safety-critical systems are to be designed for extremely rare failure events. Accurate statistical quantification of these events associated with a very low probability of occurrence requires millions of evaluations of the limit state function (LSF) involving computationally expensive numerical simulations. Variance reduction techniques like importance sampling (IS) reduce such repetitions to a few thousand. The use of a data-centric metamodel can further cut it down to a few hundred. In data-centric metamodeling approaches, the actual complex numerical analysis is performed at a few points to train the metamodel for approximating the structural response. On the other hand, a physics-informed neural network (PINN) can predict the structural response based on the governing differential equation describing the physics of the problem, without a single evaluation of the complex numerical solver, i.e., data-free. However, the existing PINN models for reliability analysis have been effective only in estimating a large range of failure probabilities (10<sup>−1</sup>∼10<sup>−3</sup>). To address this issue, the present study develops a PINN-based data-free reliability analysis for low failure probabilities (<10<sup>−5</sup>). In doing so, a two-stage PINN integrated with IS (PINN-IS) is proposed. The first stage is employed to approximate the most probable failure point (MPP) appropriately while the second stage enhances the accuracy of LSF predictions at the IS population centred on the approximated MPP. The effectiveness of the proposed approach is numerically illustrated by three structural reliability analysis examples.</div></div>","PeriodicalId":54583,"journal":{"name":"Probabilistic Engineering Mechanics","volume":"78 ","pages":"Article 103701"},"PeriodicalIF":3.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142442719","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A joint time–space extrapolation approach within the Wiener path integral technique for efficient stochastic response determination of nonlinear systems 用于高效确定非线性系统随机响应的维纳路径积分技术中的时空联合外推法

IF 3 3区工程技术 Q2 ENGINEERING, MECHANICAL

Probabilistic Engineering Mechanics

Pub Date : 2024-09-24 DOI: 10.1016/j.probengmech.2024.103685

Ilias G. Mavromatis, Ioannis A. Kougioumtzoglou

A joint time–space extrapolation approach within the Wiener path integral (WPI) technique is developed for determining, efficiently and accurately, the non-stationary stochastic response of diverse nonlinear dynamical systems. The approach can be construed as an extension of a recently developed space-domain extrapolation scheme to account also for the temporal dimension. Specifically, based on a variational principle, the WPI technique yields a boundary value problem (BVP) to be solved for determining a most probable path corresponding to specific final boundary conditions. Further, the most probable path is used for evaluating, approximately, a point of the system response joint probability density function (PDF) corresponding to a specific time instant. Remarkably, the BVP exhibits two unique features that are exploited in this paper for developing an efficient joint time–space extrapolation approach. First, the BVPs corresponding to two neighboring grid points in the spatial domain of the response PDF not only share the same equations, but also the boundary conditions differ only slightly. Second, information inherent in the time-history of an already determined most probable path can be used for evaluating points of the response PDF corresponding to arbitrary time instants, without the need for solving additional BVPs. In a nutshell, relying on the aforementioned unique and advantageous features of the WPI-based BVP, the complete non-stationary response joint PDF is determined, first, by calculating numerically a relatively small number of PDF points, and second, by extrapolating in the joint time–space domain at practically zero additional computational cost. Compared to a standard brute-force implementation of the WPI technique, the developed extrapolation approach reduces the associated computational cost by several orders of magnitude. Two numerical examples relating to an oscillator with asymmetric nonlinearities and fractional derivative elements, and to a nonlinear structure under combined stochastic and deterministic periodic loading are considered for demonstrating the reliability of the extrapolation approach. Juxtapositions with pertinent Monte Carlo simulation data are included as well.

在维纳路径积分（WPI）技术中开发了一种时空联合外推法，用于高效、准确地确定各种非线性动力系统的非稳态随机响应。该方法可视为最近开发的空域外推方案的扩展，也考虑了时间维度。具体来说，基于变异原理，WPI 技术产生了一个边界值问题 (BVP)，用于确定与特定最终边界条件相对应的最可能路径。此外，最可能路径还可用于近似评估与特定时间瞬间相对应的系统响应联合概率密度函数 (PDF) 的一个点。值得注意的是，BVP 具有两个独特的特征，本文利用这两个特征开发了一种高效的联合时空外推方法。首先，响应 PDF 空间域中两个相邻网格点对应的 BVP 不仅方程相同，而且边界条件也只有细微差别。其次，已经确定的最可能路径的时间历史固有信息可用于评估响应 PDF 中对应任意时间时刻的点，而无需求解额外的 BVP。简而言之，依靠上述基于 WPI 的 BVP 的独特优势，首先通过数值计算相对较少的 PDF 点，其次通过在联合时空域进行外推，就能确定完整的非稳态响应联合 PDF，而额外的计算成本几乎为零。与 WPI 技术的标准强制执行相比，所开发的外推方法将相关计算成本降低了几个数量级。为了证明外推法的可靠性，我们考虑了两个与具有非对称非线性和分数导数元素的振荡器有关的数值示例，以及在随机和确定性周期性组合加载下的非线性结构。此外，还将相关的蒙特卡罗模拟数据并列在一起。

{"title":"A joint time–space extrapolation approach within the Wiener path integral technique for efficient stochastic response determination of nonlinear systems","authors":"Ilias G. Mavromatis, Ioannis A. Kougioumtzoglou","doi":"10.1016/j.probengmech.2024.103685","DOIUrl":"10.1016/j.probengmech.2024.103685","url":null,"abstract":"<div><div>A joint time–space extrapolation approach within the Wiener path integral (WPI) technique is developed for determining, efficiently and accurately, the non-stationary stochastic response of diverse nonlinear dynamical systems. The approach can be construed as an extension of a recently developed space-domain extrapolation scheme to account also for the temporal dimension. Specifically, based on a variational principle, the WPI technique yields a boundary value problem (BVP) to be solved for determining a most probable path corresponding to specific final boundary conditions. Further, the most probable path is used for evaluating, approximately, a point of the system response joint probability density function (PDF) corresponding to a specific time instant. Remarkably, the BVP exhibits two unique features that are exploited in this paper for developing an efficient joint time–space extrapolation approach. First, the BVPs corresponding to two neighboring grid points in the spatial domain of the response PDF not only share the same equations, but also the boundary conditions differ only slightly. Second, information inherent in the time-history of an already determined most probable path can be used for evaluating points of the response PDF corresponding to arbitrary time instants, without the need for solving additional BVPs. In a nutshell, relying on the aforementioned unique and advantageous features of the WPI-based BVP, the complete non-stationary response joint PDF is determined, first, by calculating numerically a relatively small number of PDF points, and second, by extrapolating in the joint time–space domain at practically zero additional computational cost. Compared to a standard brute-force implementation of the WPI technique, the developed extrapolation approach reduces the associated computational cost by several orders of magnitude. Two numerical examples relating to an oscillator with asymmetric nonlinearities and fractional derivative elements, and to a nonlinear structure under combined stochastic and deterministic periodic loading are considered for demonstrating the reliability of the extrapolation approach. Juxtapositions with pertinent Monte Carlo simulation data are included as well.</div></div>","PeriodicalId":54583,"journal":{"name":"Probabilistic Engineering Mechanics","volume":"78 ","pages":"Article 103685"},"PeriodicalIF":3.0,"publicationDate":"2024-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142323814","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Stability analysis of tunnel heading in clay with nonstationary random fields of undrained shear strength 粘性土中具有非稳态随机排水剪切强度场的隧道洞口稳定性分析

IF 3 3区工程技术 Q2 ENGINEERING, MECHANICAL

Probabilistic Engineering Mechanics

Pub Date : 2024-09-23 DOI: 10.1016/j.probengmech.2024.103692

Weeradetch Tanapalungkorn , Wittawat Yodsomjai , Suraparb Keawsawasvong , Thanh Son Nguyen , Weeraya Chim-Oye , Suched Likitlersuang

The stability problem of a tunnel heading in clay remains a significant challenge in geotechnical engineering. Specifically, when considering the spatial variability of the soil, the stability factor may be influenced by geographically random fields. This study investigates the effect of random fields on a probabilistic analysis of a tunnel heading in undrained clay. The study assumes that the undrained shear strength of the clay increases linearly with depth due to a strength gradient factor. The random adaptive finite element limit analysis is employed to calculate the stability numbers for tunnel headings. Nonstationary random fields with varying vertical correlation lengths are simulated using Monte Carlo simulation technique. The stability analysis takes into account geometry parameters (i.e., cover depth ratio) and nonstationary random field of undrained shear strength parameters. (i.e., strength gradient, coefficient of variation, and vertical correlation length). The results of tunnel face stability using random adaptive finite element limit analysis have also been utilised to assess the probability of design failure over a practical range of deterministic factors of safety. In the context of probabilistic failure analysis, the failure mechanism resulting from varying vertical correlation lengths could influence the probability of design failure. The findings of this study can be of significant interest to tunnel engineering practitioners during the design phase of tunnel heading projects.

粘土隧道洞口的稳定性问题仍然是岩土工程中的一项重大挑战。具体来说，当考虑到土壤的空间变化时，稳定系数可能会受到地理随机场的影响。本研究探讨了随机场对未排水粘土中隧道走向概率分析的影响。研究假设，由于强度梯度因素，粘土的排水剪切强度随深度线性增加。采用随机自适应有限元极限分析来计算隧道洞口的稳定系数。使用蒙特卡罗模拟技术模拟了具有不同垂直相关长度的非稳态随机场。稳定性分析考虑了几何参数（即覆盖深度比）和非稳定随机场的排水抗剪强度参数。(即强度梯度、变异系数和垂直相关长度）。使用随机自适应有限元极限分析法得出的隧道面稳定性结果还被用于评估在确定性安全系数的实际范围内设计失效的概率。在概率失效分析中，不同垂直相关长度导致的失效机制可能会影响设计失效的概率。这项研究的结果对隧道工程从业人员在隧道标题项目设计阶段具有重要意义。

{"title":"Stability analysis of tunnel heading in clay with nonstationary random fields of undrained shear strength","authors":"Weeradetch Tanapalungkorn , Wittawat Yodsomjai , Suraparb Keawsawasvong , Thanh Son Nguyen , Weeraya Chim-Oye , Suched Likitlersuang","doi":"10.1016/j.probengmech.2024.103692","DOIUrl":"10.1016/j.probengmech.2024.103692","url":null,"abstract":"<div><div>The stability problem of a tunnel heading in clay remains a significant challenge in geotechnical engineering. Specifically, when considering the spatial variability of the soil, the stability factor may be influenced by geographically random fields. This study investigates the effect of random fields on a probabilistic analysis of a tunnel heading in undrained clay. The study assumes that the undrained shear strength of the clay increases linearly with depth due to a strength gradient factor. The random adaptive finite element limit analysis is employed to calculate the stability numbers for tunnel headings. Nonstationary random fields with varying vertical correlation lengths are simulated using Monte Carlo simulation technique. The stability analysis takes into account geometry parameters (i.e., cover depth ratio) and nonstationary random field of undrained shear strength parameters. (i.e., strength gradient, coefficient of variation, and vertical correlation length). The results of tunnel face stability using random adaptive finite element limit analysis have also been utilised to assess the probability of design failure over a practical range of deterministic factors of safety. In the context of probabilistic failure analysis, the failure mechanism resulting from varying vertical correlation lengths could influence the probability of design failure. The findings of this study can be of significant interest to tunnel engineering practitioners during the design phase of tunnel heading projects.</div></div>","PeriodicalId":54583,"journal":{"name":"Probabilistic Engineering Mechanics","volume":"78 ","pages":"Article 103692"},"PeriodicalIF":3.0,"publicationDate":"2024-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142316043","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Novel Bayesian updating based interpolation method for estimating failure probability function in the presence of random-interval uncertainty 基于贝叶斯更新的新插值法，用于在存在随机间隔不确定性的情况下估计故障概率函数

IF 3 3区工程技术 Q2 ENGINEERING, MECHANICAL

Probabilistic Engineering Mechanics

Pub Date : 2024-09-21 DOI: 10.1016/j.probengmech.2024.103694

Yuhua Yan , Zhenzhou Lu

Under random-interval uncertainty, the failure probability function (FPF) represents the failure probability variation as a function of the random input distribution parameter. To quickly capture the effect of the distribution parameters on failure probability and decouple the reliability-based design optimization, a novel Bayesian updating method is proposed to efficiently estimate the FPF. In the proposed method, the prior augmented failure probability (AFP) is first estimated in the space spanned by random input and distribution parameter vectors. Subsequently, by treating the distribution parameter realization as an observation, the FPF can be estimated using posterior AFP based on Bayesian updating. The main novelty of this study is the elaborate treatment of the distribution parameter realization as an observation, whereby the FPF is transformed into the posterior AFP based on Bayesian updating, and can be obtained by sharing the prior AFP simulation samples. The computational cost of the proposed method is the same as that of estimating the prior AFP. To improve the efficiency of recognizing the sample state, and improve AFP and in turn FPF estimation, the adaptive Kriging model for random-interval uncertainty was inserted into the proposed method. The feasibility and novelty of the proposed method were verified on several examples.

在随机间隔不确定性条件下，故障概率函数（FPF）表示故障概率变化与随机输入分布参数的函数关系。为了快速捕捉分布参数对失效概率的影响，并将基于可靠性的设计优化解耦，提出了一种新型贝叶斯更新方法来有效估计 FPF。在所提出的方法中，首先在随机输入和分布参数向量所跨越的空间中估计先验增强失效概率（AFP）。随后，通过将分布参数的实现视为观测值，可以使用基于贝叶斯更新的后验失效概率估计 FPF。本研究的主要新颖之处在于将分布参数实现作为一个观测点进行了精心处理，从而将 FPF 转化为基于贝叶斯更新的后验 AFP，并可通过共享先验 AFP 模拟样本获得 FPF。拟议方法的计算成本与估计先验 AFP 的计算成本相同。为了提高识别样本状态的效率，改进 AFP 并进而改进 FPF 估计，建议方法中加入了随机区间不确定性的自适应克里金模型。该方法的可行性和新颖性在多个实例中得到了验证。

{"title":"Novel Bayesian updating based interpolation method for estimating failure probability function in the presence of random-interval uncertainty","authors":"Yuhua Yan , Zhenzhou Lu","doi":"10.1016/j.probengmech.2024.103694","DOIUrl":"10.1016/j.probengmech.2024.103694","url":null,"abstract":"<div><div>Under random-interval uncertainty, the failure probability function (FPF) represents the failure probability variation as a function of the random input distribution parameter. To quickly capture the effect of the distribution parameters on failure probability and decouple the reliability-based design optimization, a novel Bayesian updating method is proposed to efficiently estimate the FPF. In the proposed method, the prior augmented failure probability (AFP) is first estimated in the space spanned by random input and distribution parameter vectors. Subsequently, by treating the distribution parameter realization as an observation, the FPF can be estimated using posterior AFP based on Bayesian updating. The main novelty of this study is the elaborate treatment of the distribution parameter realization as an observation, whereby the FPF is transformed into the posterior AFP based on Bayesian updating, and can be obtained by sharing the prior AFP simulation samples. The computational cost of the proposed method is the same as that of estimating the prior AFP. To improve the efficiency of recognizing the sample state, and improve AFP and in turn FPF estimation, the adaptive Kriging model for random-interval uncertainty was inserted into the proposed method. The feasibility and novelty of the proposed method were verified on several examples.</div></div>","PeriodicalId":54583,"journal":{"name":"Probabilistic Engineering Mechanics","volume":"78 ","pages":"Article 103694"},"PeriodicalIF":3.0,"publicationDate":"2024-09-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142323815","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Compilation of wheel-rail comprehensive irregularity spectrum for subway vehicle 地铁车辆轮轨综合不规则频谱的编制

IF 3 3区工程技术 Q2 ENGINEERING, MECHANICAL

Probabilistic Engineering Mechanics

Pub Date : 2024-09-21 DOI: 10.1016/j.probengmech.2024.103691

Qiushi Wang , Hui Zhao , Dao Gong , Jinlong Qiu , Pengfei Wu , Xiaoming Li , Xiyang Zhu , Hongyi Xiang , Tengfei Wang , Zhongmin Xiao , Jinsong Zhou

The irregularity excitation experienced by subway vehicles is mainly the result of the interaction between the track and wheel. However, in the early system design and simulation analysis of subway vehicles, most only used the traditional standard track irregularity spectrum as the input excitation, ignoring or underestimating the contribution of the wheel irregularity. Based on our statistical analysis of 200 000 km of tracking test data of subway vehicle wheel irregularities, we found that the short-wave irregularity caused by the wheels far exceeds the traditional standard track irregularity. The service condition of the vehicle is seriously affected, especially in the final stage of a wheel re-profile period. To address the above issues: Firstly, the sensitive wavelength range (16. 67–2500 mm) of subway vehicles was derived based on the axle box acceleration spectrum of IEC61373: 2010, which was very close to the wavelength range (50–2627 mm) of the wheel irregularity spectrum proposed later, demonstrating the importance of compiling a wheel irregularity spectrum; Secondly, based on the large number of tracking test data of wheel out-of-roundness, a calculation method of the wheel irregularity quantile spectrum under the Johnson non-normal transformation system was proposed; Thirdly, according to the different stages of the wheel re-profile period, the wheel irregularity spectrum is introduced to correct the short-wave segments of the traditional standard track irregularity spectrum to compile a wheel-rail comprehensive irregularity spectrum.

地铁车辆所经历的不规则激励主要是轨道与车轮之间相互作用的结果。然而，在早期的地铁车辆系统设计和仿真分析中，大多只采用传统的标准轨道不规则频谱作为输入激励，忽略或低估了车轮不规则的贡献。根据我们对 20 万公里地铁车辆车轮不平顺性跟踪测试数据的统计分析，我们发现由车轮引起的短波不平顺性远远超过了传统的标准轨道不平顺性。特别是在车轮重廓期的最后阶段，车辆的使用状况会受到严重影响。为解决上述问题：首先，敏感波长范围（16.67-2500 mm）：2010，与后来提出的车轮不规则度频谱的波长范围（50-2627 mm）非常接近，说明了编制车轮不规则度频谱的重要性；其次，基于大量的车轮失圆度跟踪测试数据，提出了约翰逊非正则变换系统下车轮不规则度量子频谱的计算方法；第三，根据车轮重廓期的不同阶段，引入车轮不规则度频谱对传统标准轨道不规则度频谱的短波段进行修正，编制轮轨综合不规则度频谱。

{"title":"Compilation of wheel-rail comprehensive irregularity spectrum for subway vehicle","authors":"Qiushi Wang , Hui Zhao , Dao Gong , Jinlong Qiu , Pengfei Wu , Xiaoming Li , Xiyang Zhu , Hongyi Xiang , Tengfei Wang , Zhongmin Xiao , Jinsong Zhou","doi":"10.1016/j.probengmech.2024.103691","DOIUrl":"10.1016/j.probengmech.2024.103691","url":null,"abstract":"<div><div>The irregularity excitation experienced by subway vehicles is mainly the result of the interaction between the track and wheel. However, in the early system design and simulation analysis of subway vehicles, most only used the traditional standard track irregularity spectrum as the input excitation, ignoring or underestimating the contribution of the wheel irregularity. Based on our statistical analysis of 200 000 km of tracking test data of subway vehicle wheel irregularities, we found that the short-wave irregularity caused by the wheels far exceeds the traditional standard track irregularity. The service condition of the vehicle is seriously affected, especially in the final stage of a wheel re-profile period. To address the above issues: Firstly, the sensitive wavelength range (16. 67–2500 mm) of subway vehicles was derived based on the axle box acceleration spectrum of IEC61373: 2010, which was very close to the wavelength range (50–2627 mm) of the wheel irregularity spectrum proposed later, demonstrating the importance of compiling a wheel irregularity spectrum; Secondly, based on the large number of tracking test data of wheel out-of-roundness, a calculation method of the wheel irregularity quantile spectrum under the Johnson non-normal transformation system was proposed; Thirdly, according to the different stages of the wheel re-profile period, the wheel irregularity spectrum is introduced to correct the short-wave segments of the traditional standard track irregularity spectrum to compile a wheel-rail comprehensive irregularity spectrum.</div></div>","PeriodicalId":54583,"journal":{"name":"Probabilistic Engineering Mechanics","volume":"78 ","pages":"Article 103691"},"PeriodicalIF":3.0,"publicationDate":"2024-09-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142316044","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0