Robust Variance Reduction Techniques for Improving Monte Carlo Simulation in Structural Reliability Analyses.
Robust Variance Reduction Techniques for Improving Monte Carlo Simulation in Structural Reliability Analyses.
Alireza Hajishabanian1 Karim Laknejadi2 Panam Zarfam3
1) Student of science and university of tehran Email:
2) Professor assistant of science and university of tehran Email:
3) Professor assistant of science and university of tehran Email:
IOI Code
:
XBFC-DEFBA
Publication :
4th International Congress on Engineering, Technology and Applied Science - Auckland University of Technology(etas2019.com)
Abstract :
Monte Carlo (MC) simulation is recognized as a robust mathematical method for determining the statistical properties of uncertain variables and estimating the probability of special events related to the performance of structures. Beside several advantages of MC simulation for studying uncertainties, applying this method is usually faced with one computational challenge. For reaching to more accurate results in this simulation, we are usually perforce to employ more replication and consequently taking large samples of stochastic variables. This problem sometimes increases the computational cost of simulation and run time of analysis as far as the researcher will be forced to search for the other reliability methods to simulate the effects of uncertainties. In this paper four robust variance reduction techniques (VRTs) are investigated for improving the efficiency of MC simulation in structural reliability analyses and decreasing the computational cost of simulation by using targeted calculations. Accordingly, four hybrid combinations of these techniques are introduced and investigated to demonstrate their overall performance. For this purpose, we have applied these variance reduction techniques in two case studies of structural problems. The obtained results represented that these numerical techniques outperform other well-known methods in the literature.
Keywords : monte carlo simulation uncertainties variance reduction techniques reliability analyses