Sampling-based methods are commonly used to propagate uncertainty through models for complex systems (Helton and Davis (2003), Helton et al. (2006)). Replicated sampling involves repeating a sampling-based uncertainty propagation for several independent samples of the same size (Iman (1982)). Variance between replicates indicates the numerical uncertainty in analysis results that derives from the sampling-based method. Results from the replicates can be used to estimate confidence intervals for analysis results and to determine whether the sample size in use is sufficient to obtain statistically stable results. Replicated sampling can be used to assess the adequacy of the sample size in situations where more formal statistical procedures are not applicable (Iman (1982)).
Hansen, C. W., Helton, J. C., & Sallaberry, C. J. (2010). Use of replicated Latin hypercube sampling to estimate sampling variance in uncertainty and sensitivity analysis results for the geologic disposal of radioactive waste. In Procedia - Social and Behavioral Sciences (Vol. 2, pp. 7674–7675). https://doi.org/10.1016/j.sbspro.2010.05.173