Two partitioning methods for multi-area studies in large power systems
2015 (English)In: International Transactions on Electrical Energy Systems, E-ISSN 2050-7038, Vol. 25, no 4, 648-660 p.Article in journal (Refereed) Published
Multi-area studies are an important tool for today's and future power systems. In this paper, a two-step algorithm for creating multi-area models is presented that, first, identifies areas, and, second, computes reduced models of these areas. For the first step, two new methods to identify areas in power systems have been developed. The first method is based upon spectral partitioning, whereas the second one is formulated as a linear optimization problem. The methods are compared in terms of computation time on the IEEE 118 bus system, and the first method clearly stands out in this comparison. The first method is then applied to the IEEE 300 bus system and to a model of the Polish power system with 2746 buses to study how it scales in large power systems. Even in the latter case, it runs in less than 30s. For the second step, existing equivalencing methods can be used. As an example, radial, equivalent, and independent equivalents are used to model the areas identified by the partitioning methods. The partitioning and equivalencing methods have been tested on the IEEE 118 bus system by running 1000 regular and optimal power flows. Comparisons with the original IEEE 118 bus system in terms of flows, costs and losses are carried out.
Place, publisher, year, edition, pages
2015. Vol. 25, no 4, 648-660 p.
spectral partitioning, power system partitioning, multi-area study, REI
IdentifiersURN: urn:nbn:se:kth:diva-158949DOI: 10.1002/etep.1864ISI: 000353388700006ScopusID: 2-s2.0-84928023607OAI: oai:DiVA.org:kth-158949DiVA: diva2:781356
QC 201508052015-01-162015-01-162015-08-05Bibliographically approved