A hybrid algorithm for Monte Carlo/enumeration based composite reliability evaluation [power systems]

The authors discuss possible hybrid schemes for combining enumeration and Monte Carlo approaches in composite reliability evaluation. One intuitive approach is to use enumeration to select transmission contingencies, followed by Monte Carlo sampling for the generation contingencies. It is shown that this technique can be seen as a stratified sampling scheme. Computational experiments with a 124-bus system derived from a Brazilian configuration indicated that this hybrid scheme was not attractive, even if the optimal strata allocation and optimal ranking schemes were used. The basic reason seems to be that the contributions of transmission contingencies to composite reliability indices, when combined with generation contingencies, are not concentrated in a few `severe´ outages, but are `spread out´ in a more or less uniform way. Further experiments with other systems are required before these conclusions are generalized. Finally, it is pointed out that the so-called variance reduction schemes can be seen as hybrid schemes