An adaptive ensemble of surrogate models based on hybrid measure for reliability analysis

2021 ◽  
Vol 65 (1) ◽  
Author(s):  
Changcong Zhou ◽  
Hanlin Zhang ◽  
Qi Chang ◽  
Xiaokang Song ◽  
Chen Li
Keyword(s):  
Reliability Analysis ◽  
Surrogate Models

Shallow and Deep Artificial Neural Networks for Structural Reliability Analysis

2020 ◽  
Author(s):  
Wellison J. S. Gomes
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Reliability Analysis ◽  
Structural Reliability ◽  
Deep Neural Networks ◽  
Surrogate Models ◽  
The Other ◽  
Computational Costs ◽  
Structural Reliability Analysis ◽  
Artificial Neural

Abstract Surrogate models are efficient tools which have been successfully applied in structural reliability analysis, as an attempt to keep the computational costs acceptable. Among the surrogate models available in the literature, Artificial Neural Networks (ANNs) have been attracting research interest for many years. However, the ANNs used in structural reliability analysis are usually the shallow ones, based on an architecture consisting of neurons organized in three layers, the so-called input, hidden and output layers. On the other hand, with the advent of deep learning, ANNs with one input, one output, and several hidden layers, known as deep neural networks, have been increasingly applied in engineering and other areas. Considering that many recent publications have shown advantages of deep over shallow ANNs, the present paper aims at comparing these types of neural networks in the context of structural reliability. By applying shallow and deep ANNs in the solution of four benchmark structural reliability problems from the literature, employing Monte Carlo simulation and adaptive experimental designs, it is shown that, although good results are obtained for both types of ANNs, deep ANNs usually outperform the shallow ones.

On the Use of Surrogate Models in Reliability-Based Analysis of Dented Pipes

2016 ◽  
Author(s):  
Sherif Hassanien ◽  
Muntaseer Kainat ◽  
Samer Adeeb ◽  
Doug Langer
Keyword(s):  
Finite Element ◽  
Reliability Analysis ◽  
Random Systems ◽  
Probability Of Failure ◽  
Surrogate Models ◽  
Operating Conditions ◽  
Additional Stress ◽  
Internal Stresses ◽  
Stress Risers ◽  
Reliability Method

Pipeline dents lead to changes in the stress/strain state of the pipe body, making it more susceptible to integrity concerns. This susceptibility is especially prevalent in cases where additional stress risers such as crack and/or corrosion features interact with the dented region. While some guidance is available in codes, regulations, and industry best practices, there is substantial room for innovation and improvement to ensure pipeline safety. Existing explicit models are primarily based on experimental correlations and historical findings using simple parameters such as dent depth and location on the pipeline. Moreover, these models are subjected to a substantial uncertainty in both accuracy and precision. This paper presents a state-of-the-art methodology for analyzing dents and dents associated with stress risers through the use of finite element method (FEM) as a mechanical model and reliability analysis to address uncertainties associated with input variables. FEM is used to model the full geometry of dents and any interacting stress risers reported by inline inspection (ILI) to be incorporated into calculations of the internal stresses/strains within the feature. Theoretically, FEM and reliability analysis can be integrated through reliability-based stochastic finite element methodologies due to the absence of closed form mechanical models of dented pipes. However, these methodologies are computationally prohibitive and not suited/designed for frequent integrity analysis. This study aims at further advancing such integration by combining FEM with reliability science to account for pipe properties and measurement uncertainties in order to determine the probability of failure under different operating conditions using surrogate models. This provides the opportunity to more accurately assess the risk posed by ILI reported dent features. Herein, surrogate models refer to the response surface method (RSM) which is considered as a valuable tool for obtaining insight into the behavior of structural random systems at low computational costs. The proposed approach was applied focusing on a plain dent, a dent interacting with a corrosion feature, and a dent interacting with a crack feature. First Order Reliability Method (FORM) is used to evaluate the probability of failure/reliability using the resulting RSM non-linear limit states for each dent feature.

AN ITERATIVE HDMR APPROACH FOR ENGINEERING RELIABILITY ANALYSIS

2021 ◽  
Vol 8 (1) ◽  
Author(s):  
Qian Wang
Keyword(s):  
Reliability Analysis ◽  
Probabilistic Analysis ◽  
Surrogate Models ◽  
Second Order ◽  
High Dimensional Model Representation ◽  
Performance Function ◽  
Reliability Indices ◽  
Reliability Method ◽  
Sample Points ◽  
A Performance

Engineering reliability analysis has long been an active research area. Surrogate models, or metamodels, are approximate models that can be created to replace implicit performance functions in the probabilistic analysis of engineering systems. Traditional 1st-order or second-order high dimensional model representation (HDMR) methods are shown to construct accurate surrogate models of response functions in an engineering reliability analysis. Although very efficient and easy to implement, 1st-order HDMR models may not be accurate, since the cross-effects of variables are neglected. Second-order HDMR models are more accurate; however they are more complicated to implement. Moreover, they require much more sample points, i.e., finite element (FE) simulations, if FE analyses are employed to compute values of a performance function. In this work, a new probabilistic analysis approach combining iterative HDMR and a first-order reliability method (FORM) is investigated. Once a performance function is replaced by a 1st-order HDMR model, an alternate FORM is applied. In order to include higher-order contributions, additional sample points are generated and HDMR models are updated, before FORM is reapplied. The analysis iteration continues until the reliability index converges. The novelty of the proposed iterative strategy is that it greatly improves the efficiency of the numerical algorithm. As numerical examples, two engineering problems are studied and reliability analyses are performed. Reliability indices are obtained within a few iterations, and they are found to have a good accuracy. The proposed method using iterative HDMR and FORM provides a useful tool for practical engineering applications.

Sign in / Sign up

Export Citation Format

Share Document