Hong-Shuang Li

Application of subset simulation in reliability estimation of underground pipelines

This paper presents a computational framework for implementing an advanced Monte Carlo simulation method, called Subset Simulation (SS) for time-dependent reliability prediction of underground flexible pipelines. The SS can provide better resolution for low failure probability level of rare failure events which are commonly encountered in pipeline engineering applications. Random samples of statistical variables are generated efficiently and used for computing probabilistic reliability model. It gains its efficiency by expressing a small probability event as a product of a sequence of intermediate events with larger conditional probabilities. The efficiency of SS has been demonstrated by numerical studies and attention in this work is devoted to scrutinise the robustness of the SS application in pipe reliability assessment and compared with direct Monte Carlo simulation (MCS) method. Reliability of a buried flexible steel pipe with time-dependent failure modes, namely, corrosion induced deflection, buckling, wall thrust and bending stress has been assessed in this study. The analysis indicates that corrosion induced excessive deflection is the most critical failure event whereas buckling is the least susceptible during the whole service life of the pipe. The study also shows that SS is robust method to estimate the reliability of buried pipelines and it is more efficient than MCS, especially in small failure probability prediction.

Structural reliability analysis of multiple limit state functions using multi-input multi-output support vector machine

Selecting and using an appropriate structural reliability method is critical for the success of structural reliability analysis and reliability-based design optimization. However, most of existing structural reliability methods are developed and designed for a single limit state function and few methods can be used to simultaneously handle multiple limit state functions in a structural system when the failure probability of each limit state function is of interest, for example, in a reliability-based design optimization loop. This article presents a new method for structural reliability analysis with multiple limit state functions using support vector machine technique. A sole support vector machine surrogate model for all limit state functions is constructed by a multi-input multi-output support vector machine algorithm. Furthermore, this multi-input multi-output support vector machine surrogate model for all limit state functions is only trained from one data set with one calculation process, instead of constructing a series of standard support vector machine models which has one output only. Combining the multi-input multi-output support vector machine surrogate model with direct Monte Carlo simulation, the failure probability of the structural system as well as the failure probability of each limit state function corresponding to a failure mode in the structural system can be estimated. Two examples are used to demonstrate the accuracy and efficiency of the presented method.

Discrete Optimum Design for Truss Structures by Subset Simulation Algorithm

AbstractThis article deals with the design optimization of truss structures with discrete design variables, which remains quite a challenging task in structural design. A new discrete search strategy based on the recently developed subset simulation optimization algorithm is proposed in detail for this type of structural optimization. The discrete design variables are transformed into standard normal variable space to implement the sampling procedure in subset simulation optimization, while the optimization is processed in the discrete design space in the mean time. The performance of the proposed method is illustrated by four representative benchmark optimization problems. Comparisons are made with other well known stochastic optimization algorithms. It is found that the proposed method can produce optimum designs as good as or better than those of other stochastic optimization algorithms.

Hybrid Dimension-Reduction Method for Robust Design Optimization

Robust design optimization has gained increasing concern in the engineering design process because it can provide an economical design that is insensitive to variations in the input variables without eliminating their causes. Robustness assessment, which estimates the mean values and standard deviations of the objective function and constraint functions, is an important and inevitable component in robust design optimization. A hybrid dimension-reduction method is proposed for efficient and accurate robustness assessment in this paper. It is a combination of the univariate dimension-reduction method and the bivariate dimension-reduction method for the balance of efficiency and accuracy. The significant random variables are identified by their variation contributions to the output variation of the objective function and constraints using the univariate dimension-reduction method. The approximate part of the significant random variables in the approximate function is then extended to use the bivariate dimens...

Identifying the Probability Distribution of Fatigue Life Using the Maximum Entropy Principle

It is well-known that the fatigue lives of materials and structures have a considerable amount of scatter and they are commonly suggested to be considered in engineering design. In order to reduce the introduction of subjective uncertainties and obtain rational probability distributions, a computational method based on the maximum entropy principle is proposed for identifying the probability distribution of fatigue life in this paper. The first four statistical moments of fatigue life are involved to formulate constraints in the maximum entropy principle optimization problem. An accurate algorithm is also presented to find the Lagrange multipliers in the maximum entropy distribution, which can avoid the numerical singularity when solving a system of equations. Two fit indexes are used to measure the goodness-of-fit of the proposed method. The rationality and effectiveness of the proposed method are demonstrated by two groups of fatigue data sets available in the literature. Comparisons among the proposed method, the lognormal distribution and the three-parameter Weibull distribution are also carried out for the investigated groups of fatigue data sets.

Uncertainty propagation in fatigue crack growth analysis using dimension reduction technique

This paper proposes a general probabilistic methodology for uncertainty propagation in fatigue crack growth analysis under both constant amplitude and variable amplitude loadings. A recently developed small time scale model is used to predict the deterministic fatigue crack growth curve (a-N curve). The dimension reduction technique is used for uncertainty propagation in fatigue crack growth analysis. The basic idea is to avoid direct simulations and focuses on the statistical moment behaviour of output random variables. A modified Chebyshev algorithm is presented to improve the approximation accuracy when calculating the integral points and associated weights for an arbitrary probabilistic distribution. Uncertainties of some material properties are considered as input random variables and propagated through the mechanical model. Prediction result of the proposed methodology is compared with the direct Monte Carlo Simulation (MCS) and is verified using experimental data of aluminium alloys under constant amplitude loading and block loading.

Combined size and shape optimization of truss structures using subset simulation optimization

This paper presents an approach to solve the combined size and shape design optimization problems using recently developed subset simulation optimization for both continuous and discrete design variables. Except for the componentwise Metropolis–Hasting algorithm, a recently developed adaptive conditional sampling algorithm is also employed as an alternative approach for generating new conditional samples (candidate designs) for each simulation level, which enhances the accuracy and stability of the optimization process. Besides, a double-criterion sorting algorithm is used to handle the design constraints and integrate them in the generation of conditional samples during the Markov Chain Monte Carlo simulation, and the inverse transform method is employed to deal with the discrete design variables. Totally, four numerical examples are considered, including a 15-bar 2D truss, an 18-bar 2D truss, a 39-bar 3D truss and a truss-type landing gear of an unmanned aerial vehicle. The optimal designs obtained from subset simulation optimization using either the componentwise Metropolis–Hasting algorithm or the adaptive conditional sampling algorithm succeed in substantially reducing the weights of the truss-type structures under design constraints in terms of the member stress, the Euler buckling and the nodal displacement. The computational results indicate the proposed method can be taken as an alternative tool for structural optimization design on truss structures when involving the combined size and shape design.

SUBSET SIMULATION FOR ASSESSING STRUCTURAL RELIABILITY OF MULTIPLE LIMIT STATE FUNCTIONS

It remains a challenging task to calculate failure probabilities of multiple limit state functions (LSFs) using a single run of Subset Simulation in structural reliability analysis. To address this issue, this article presents a variant of standard Subset Simulation (SS), in which a unified intermediate event is designed to drive the simulation procedure progressively approaching multiple failure regions defined by all LSFs simultaneously. All failure probabilities of multiple LSFs are obtained by a single run of SS, which bypasses the sorting difficulty arising from the multiple LSFs. A representative example is used to demonstrate the efficiency, accuracy and robustness of the presented SS method.