Pregled bibliografske jedinice broj: 410828
Binary Decision Tree Development for Probabilistic Safety Assessment Applications
Binary Decision Tree Development for Probabilistic Safety Assessment Applications // Proceedings of the 7th International Conference on Nuclear Option in Countries with Small and Medium Electricity Grids / Čavlina, Nikola ; Pevec, Dubravko ; Bajs, Tomislav (ur.).
Zagreb: Hrvatsko nuklearno društvo, 2008. str. S-051011 (predavanje, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)
CROSBI ID: 410828 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Binary Decision Tree Development for Probabilistic
Safety Assessment Applications
Autori
Šimić, Zdenko ; Banov, Reni ; Mikuličić, Vladimir
Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni
Izvornik
Proceedings of the 7th International Conference on Nuclear Option in Countries with Small and Medium Electricity Grids
/ Čavlina, Nikola ; Pevec, Dubravko ; Bajs, Tomislav - Zagreb : Hrvatsko nuklearno društvo, 2008, S-051011
ISBN
978-953-55224-0-9
Skup
7th International Conference on Nuclear Option in Countries with Small and Medium Electricity Grids
Mjesto i datum
Dubrovnik, Hrvatska, 25.05.2008. - 29.05.2008
Vrsta sudjelovanja
Predavanje
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
BDD ; PSA ; PRA
Sažetak
The aim of this article is to describe state of the development for the relatively new approach in the probabilistic safety analysis (PSA). This approach is based on the application of binary decision diagrams (BDD) representation for the logical function on the quantitative and qualitative analysis of complex systems that are presented by fault trees and event trees in the PSA applied for the nuclear power plants risk determination. Even BDD approach offers full solution comparing to the partial one from the conventional quantification approach there are still problems to be solved before new approach could be fully implemented. Major problem with full application of BDD is difficulty of getting any solution for the PSA models of certain complexity. This paper is comparing two approaches in PSA quantification. Major focus of the paper is description of in-house developed BDD application with implementation of the original algorithms. Resulting number of nodes required to represent the BDD is extremely sensitive to the chosen order of variables (i.e., basic events in PSA). The problem of finding an optimal order of variables that form the BDD falls under the class of NP- complete complexity. This paper presents an original approach to the problem of finding the initial order of variables utilized for the BDD construction by various dynamical reordering schemes. Main advantage of this approach compared to the known methods of finding the initial order is with better results in respect to the required working memory and time needed to finish the BDD construction. Developed method is compared against results from well known methods such as depth- first, breadth-first search procedures. Described method may be applied in finding of an initial order for fault trees/event trees being created from basic events by means of logical operations (e.g. negation, and, or, exclusive or). With some testing models a significant reduction of used memory has been achieved, sometimes even more than twice. Such significant reductions have been achieved by means of a slight increase of time necessary for the creation of the initial order. This is important result because the time for the order determination is negligible compared to the time needed for the BDD construction. Finally it has to be noted that this approach allows further improvements which will be the subject matter of the continuing research.
Izvorni jezik
Engleski
Znanstvena područja
Elektrotehnika
POVEZANOST RADA
Projekti:
MZO-ZP-036-0361590-1589 - Nuklearne elektrane za održivu proizvodnju električne energije (Feretić, Danilo, MZO ) ( CroRIS)
Ustanove:
Fakultet elektrotehnike i računarstva, Zagreb