PARALLEL IMPLEMENTATION OF PROBABILISTIC SHORT-CIRCUIT ANALYSIS BY THE MONTE-CARLO APPROACH

This paper describes the implementation of a Monte Carlo based short-circuit analysis package in a hypercube parallel computer. The program generates accurate estimates of short-circuit current probability distribution curves which can be used in a variety of planning and design studies. Both network modeling aspects, such as network reduced equivalencing and mutual coupling representation, as well as the impact of the sample sizes used in Monte Carlo simulations are fully analyzed and illustrated by numerical results obtained testing on real-world systems. These examples clearly illustrate how computational efficiency in parallel machines varies with the ratio of communication to computation (communication overhead) and of the load balancing among the parallel computer processors.