Loading [a11y]/accessibility-menu.js
MEAK-MCS: Metamodel Error Measure Function based Active Learning Kriging with Monte Carlo Simulation for Reliability Analysis | IEEE Conference Publication | IEEE Xplore

MEAK-MCS: Metamodel Error Measure Function based Active Learning Kriging with Monte Carlo Simulation for Reliability Analysis


Abstract:

The active learning reliability analysis method combining Kriging and MCS, termed as AK-MCS, has received great attention over the past few years. When the limit-state fu...Show More

Abstract:

The active learning reliability analysis method combining Kriging and MCS, termed as AK-MCS, has received great attention over the past few years. When the limit-state function is replaced by Kriging metamodel, the estimation of failure probability will be influenced by metamodel error. In this paper, an error measure function is developed to quantify the influence of metamodel error. Then, it is introduced in the termination conditions of Kriging update in AK-MCS. A metamodel error measure function based AK-MCS (MEAK-MCS) method is finally proposed. With the error measure function, the update of Kriging can be efficiently terminated. Numerical examples are presented to test the efficiency and accuracy of MEAK-MCS. Results indicate that MEAK-MCS can more efficiently estimate the failure probability than AK-MCS and has similar accuracy with AK-MCS.
Date of Conference: 06-08 May 2019
Date Added to IEEE Xplore: 08 August 2019
ISBN Information:
Conference Location: Porto, Portugal

References

References is not available for this document.