English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

This paper develops mfEGRA, a multifidelity active learning method usingdata-driven adaptively refined surrogates for failure boundary location inreliability analysis. This work addresses the issue of prohibitive cost ofreliability analysis using Monte Carlo sampling for expensive-to-evaluatehigh-fidelity models by using cheaper-to-evaluate approximations of thehigh-fidelity model. The method builds on the Efficient Global ReliabilityAnalysis (EGRA) method, which is a surrogate-based method that uses adaptivesampling for refining Gaussian process surrogates for failure boundary locationusing a single-fidelity model. Our method introduces a two-stage adaptivesampling criterion that uses a multifidelity Gaussian process surrogate toleverage multiple information sources with different fidelities. The methodcombines expected feasibility criterion from EGRA with one-step lookaheadinformation gain to refine the surrogate around the failure boundary. Thecomputational savings from mfEGRA depends on the discrepancy between thedifferent models, and the relative cost of evaluating the different models ascompared to the high-fidelity model. We show that accurate estimation ofreliability using mfEGRA leads to computational savings of $\sim$46% for ananalytic multimodal test problem and 24% for a three-dimensional acoustic hornproblem, when compared to single-fidelity EGRA. We also show the effect ofusing a priori drawn Monte Carlo samples in the implementation for the acoustichorn problem, where mfEGRA leads to computational savings of 45% for thethree-dimensional case and 48% for a rarer event four-dimensional case ascompared to single-fidelity EGRA.

mfEGRA: Multifidelity Efficient Global Reliability Analysis through Active Learning for Failure Boundary Location