A fast and efficient method for reliability evaluation of interconnected power systems-preferential decomposition method Academic Article uri icon

abstract

  • A new method for reliability evaluation of interconnected power systems is presented in this paper. In the Simultaneous Decomposition-Simulation (Sides) method, efficiency of the decomposition process decreases rapidly as the sets get decomposed into smaller probability sets. This is because, the number of states being classified as acceptable sets or identical area loss-of-load sets in generation deficient areas is small and consequently, the number of sets which are generated for a given threshold probability is large. The number of sets generated has serious implications on the memory requirements and frequency calculations which depend on the examination of subset intersection. In the new Preferential Simultaneous Decomposition-Simulation (Presides) method five concepts have been introduced which improve the efficiency of the decomposition process. A characteristic value is assigned to each of the areas for the maximum state of the set depending on excess generation or deficient gener-ation. Then the maximal flow algorithm is performed to utilize generation, first from excess generation areas and then from deficient generation areas. This increases the states in deficient generation areas being classified as acceptable sets or identical area loss-of-load sets to more than one state. Also in the decomposition of loss-of-Ioad sets, if all the areas which have deficient generation are loss-of-load, then the neighboring help to these areas is reduced to zero which increases the states in identical area loss-of-load set being generated. Application of these concepts to the decomposition process results in much fewer sets and improves the efficiency of the decomposition phase. The Presides method is shown to be 35 times faster than the Sides method. 1993 IEEE

published proceedings

  • IEEE Transactions on Power Systems

author list (cited authors)

  • Gubbala, N., & Singh, C.

citation count

  • 8

complete list of authors

  • Gubbala, N||Singh, C

publication date

  • May 1994