Reliability of Technical Systems Estimated by Enhanced Monte Carlo Simulation

Abstract

Computation of the reliability of large technical systems is usually a very di cult problem for realistic systems, and it is generally not possible to calculate the exact reliability. There are many techniques for approximate calculations, but they are often complicated and di cult to implement. In this paper the development of a new method based on Monte Carlo simulation for e cient calculation of system reliability is described. Standard Monte Carlo simulation forms a simple and robust alternative for calculating system reliability. If one can generate large samples, the law of large numbers ensures that the estimated reliability will be accurate as well. This may, however, be a very time consuming operation. The new method introduces a parametrized system that corresponds to the given system for a specific parameter value. By using regularity of the system reliability as a function of the introduced parameter, the system reliability for our original system can be predicted accurately from relatively small samples.

Advances in Simulation-Based Uncertainty Quantification and Reliability Analysis

“…Zhang and Taflanidis (2018), Moustapha et al (2018), and Sundar and Shields (2019) discuss and compare approaches for surrogate model development-most notably Kriging-and support-vector-regression-based techniques. Naess and Bo (2018), Zhang and Taflanidis (2018), and Sundar and Shields (2019) further discuss issues related to sampling for Monte Carlo simulations or surrogate model development, while Christou et al (2018) employs an optimal set of random field samples for hazard modeling. Lastly, Cai and Mahadevan (2018) discuss approaches to leverage big data for uncertainty quantification purposes.…”