In the present work, the effect of correlation between the model parameters on the probability of failure for a

concrete structure is being explored. In the usual case when the joint probability distribution function between the

correlated parameters is not analytically modelled, the derived probability of failure may contain significant

uncertainty. In order to investigate the magnitude of this type of uncertainty, different dependence types having the

same correlation coefficient are assumed. Copula functions are used for the construction of different multivariate

distributions as well as for Monte Carlo sampling. The probability of failure for a simple concrete element is

calculated and a range of feasible values for the probability of failure is obtained.